AI Workshop:
Feedback & Future Preferences
An interactive, data-driven synthesis of survey responses submitted by MBA students following the Generative AI for Business case study workshop.
SURVEY RESULTS
Quantitative Metrics & Ratings
Value to Professional Development
4.9 / 5.0 AvgLikelihood to Recommend
5.0 / 5.0 AvgWORKSHOP FEEDBACK
Hands-On Cases & Analytical "Data Traps"
Learning Through Real Cases
- Hands-on over Lecture: Students expressed strong appreciation for practical, interactive case work (the BrewPoint Coffee case study) rather than listening to passive slides.
- Functional Speed: The speed with which AI models handled and structured data was impressive, demonstrating productivity gains.
- Framing: Focus was placed on how to structure a business case problem so AI can help solve it.
Uncovering AI "Data Traps"
- Eye-Opening Debrief: Understanding where AI models confidently output incorrect business assumptions was a primary learning.
- Correlation vs. Causality: The "Staff Turnover Trap" showed that AI will recommend fixing turnover to solve revenue, failing to see that low revenue/wages actually caused the high turnover.
- Human Interrogation: Reinforced that the manager's role is to stress-test the model's logic, questioning causality and selection bias.
STRATEGIC TAKEAWAYS
Agentic Workflows & Decision Agency
Structured Agentic Workflows
- Moving Beyond Casual Chats: Students learned that chatting with AI is a poor fit for rigorous work. Complex business problems require step-by-step agentic workflows.
- Intermediate Reasoning: Breaking prompts into input, structured reasoning, testing, and feedback yields reliable, work-ready outputs.
- Tasks, Not Workflows: AI automates specific analytical tasks, but the overall workflow structure remains the domain of the manager.
Human Ownership & Verification
- Decision Assistant, Not Maker: A key consensus from the feedback was that AI is a tool to help make choices, but humans must retain final ownership.
- Model Performance Discrepancies: Observing how vastly different AI tools (Gemini, ChatGPT, Claude) and their free vs. premium models perform in practice.
- Critical Review: If you cannot explain or verify the AI's data recommendations, you cannot stand by them.
FUTURE PREFERENCES
Format Preferences for Expanded AI Offerings
Students voted on how the program should expand AI training. There is a strong appetite for deeper, structured courses.
A Full Elective Course on AI for Business
The top-ranked option. Students want a dedicated, comprehensive elective course to study advanced prompt engineering, custom agents, and AI economics in detail.
Multi-Week Workshop Series (4–6 sessions)
Students appreciate the workshop format but want sequential sessions. Note: One additional student noted in comments that they would have selected this format.
Short Embedded Sessions in Core Classes
Integrating 1–2 hour specialized AI sessions directly into existing finance, supply chain, or strategy courses to provide context-specific tool learning.
Standalone Semiannual Workshops
Continuing the current standalone format (like today's session) offered once or twice per semester as optional extracurricular sessions.
QUALITATIVE FEEDBACK