When Duane Lefevre assigned case studies in his marketing courses, he noticed a pattern. Students were feeding cases directly into AI chatbots, plugging in assignment questions and rubrics, and submitting the output as their own work.
“My discussion boards and online classes and my case studies were just of less and less value,” Lefevre says.
The fix, it turned out, was AI itself — this time designed to make students think rather than think for them.
Lefevre, a part-time lecturer of marketing at D'Amore-McKim, partnered with his husband John Biske — a 27-year Fidelity Investments veteran and founder of Support Savvy Consulting — to build an interactive case study tool that students can't game. The platform, developed using Anthropic's Claude, replaces traditional written case studies with simulated executive interviews that require students to think critically, ask follow-up questions and dig for information on their own. Northeastern's university-wide partnership with Anthropic gives all faculty, staff and students access to Claude.
How it works
The tool puts questioning skills at a premium, challenging students to draw out information from a cast of executive personas that Lefevre programmed to think, react and push back like real people.
Faculty use a case builder to create a fictional company and a set of executive personas — a founder, an operations director, a sustainability advisor, a CFO — each loaded with background knowledge, personality traits and opinions relevant to the case scenario. Students receive only a few paragraphs of company context, then must interview the personas to gather enough information to complete their assignment.
The personas also emote, physically leaning forward, sighing or removing their glasses mid-conversation, which students say makes interviews feel less like chatting with a bot and more like talking to a real person. Faculty can even program conflict between characters, creating organizational dynamics students have to navigate.
“It really feels like they're interviewing people,” Lefevre says.
Building cases from scratch
Lefevre uses Claude to help develop new case scenarios before ever opening the builder. For a recent collaboration with a faculty partner at FGV, a business school in São Paulo, he asked Claude to generate options for a case on corruption in Brazil, settled on a startup founder facing a bribery demand, and had Claude build out three character backstories, a company narrative, and a dilemma. The whole process took a single conversation.
Once a case is complete, Claude generates a teaching note — a detailed document covering the analytical frameworks and tools a faculty member might use to discuss the case — making it easy to share cases across courses and instructors.
“It's been exciting to watch the AI Case Builder Tool grow from an idea into a resource that can meaningfully support teaching and learning,” says Barbara Barsch, academic instructional technologist. “It has the potential to help faculty quickly create authentic learning experiences that promote deeper student engagement.”
What students get out of it
When students finish interviewing, the tool generates an export summary: total time spent, number of questions asked per persona, an AI-generated analysis of their questioning quality, and a full transcript of every conversation. Lefevre requires students to submit the export summary alongside their written assignment.
Lefevre has been using the tool for three semesters, and faculty at DMSB are beginning to explore it for courses ranging from international business to organizational design.
The team consulted with Kwong Chan, executive director of the D'Amore-McKim AI Strategic Hub (DASH), which works with faculty, staff, and students across research, teaching, and AI tool development.
“This is a great example of our faculty trying something new completely on their own,” Chan says. “This tool is a learning tool for students to immerse themselves in a real-world context. They've really gone outside the box.”
A different kind of classroom engagement
Larissa Marchiori Pacheco has used the tool in two of her international business and strategy courses — one built around organizational structure and decision-making, another centered on the growth strategy of a feminine hygiene product company entering new markets. In both cases, she says, the personas made the difference.
“What made the tool genuinely powerful was the personas,” Pacheco says. “The personas didn't just deliver information — they expressed personality through their language, their hesitations, their frustration when students challenged them, and their enthusiasm when a good idea landed.”
That behavioral realism, she says, is difficult to replicate with a traditional case study or an online search, which meant students couldn't prepare a canned answer in advance.
“Students had to read the room, adapt in real time and think quickly,” Pacheco says. “The tool gave them a genuine challenge: figure out how to counter an argument, how to move a reluctant stakeholder, or how to reframe a problem for someone who sees it completely differently. That's the kind of learning I want in my classroom.”
Faculty interested in trying the tool can access it through their Northeastern Claude account. Barsch's team serves as the point of contact for support.
For Lefevre, the goal was never to ban AI from the classroom. It was to make students use their own minds first.
“A student who doesn't look at the tables in a case study gets less out of it,” he says. “This is the same thing.”