AI INTERACTION
T he goal of any prepare students to be capable of making high‑stakes decisions with incomplete information in uncertain or unfavourable conditions. While frameworks offer a much-needed structure and shared language for such preparation, they can also appear detached from the ambiguity and messiness of real-world decision-making and are unlikely to reflect all scenarios or industries. Their use in teaching, therefore, requires a degree of caution to ensure that students learn when individual frameworks can be trusted and when they must be adapted or even abandoned. It is this challenge that custom-designed, ChatGPT 4.0-powered simulations can help address, by enabling students to stress-test frameworks and refine their thinking through experience rather than abstraction. MBA‑level course on strategy is to Connecting research & real-world complexity I developed five custom simulations for a course on strategic innovation at the Indian School of Business, with the specific aim of: • Bringing relevant academic research into the class and incorporating it in discussions and learning • Helping students explore, apply and adapt frameworks in dynamic scenarios • Recreating the real-world complexity of decision‑making under uncertainty. The design did not require complex coding. Instead, it involved careful testing and the creation of guardrails so that ChatGPT could simulate a realistic decision environment without prematurely revealing information. Rooted in peer-reviewed research papers, each of the five simulations presented an opportunity for the instructor to show students that by engaging with a theory and reflecting on where it holds and where it doesn’t, they can improve their decision‑making skills. They also offer a unique immersive experience, in which students assume managerial roles, make tough trade-offs and receive real-time, framework-informed feedback from an AI interface.
All too often, research papers remain distant from classroom teaching, with their insights buried under dense models. However, simulations give students a chance to interact with academic facts in an engaging and personalised learning format. They do this by embedding research logic into game dynamics and layering ChatGPT’s natural language fluency on top. For example, one of the five simulations is inspired by a research paper on commercialisation in the biotech industry, penned by Harvard Business School professor Gary Pisano. It aims to help students learn how to manage risk and navigate trade-offs in their choices. The Survive the Dominant Design simulation, meanwhile, is based on a paper from MIT Sloan School of Management’s James Utterback and Fernando Suárez that studies how the success of new enterprises and a product’s market entry is linked to the level of competition and structure of an industry ( see boxout on page 33 for more details of all five simulations ). If these research papers were given to students as isolated pre-reads, it is unlikely that many would have been able to appreciate the underlying research or connect it with a real-world context effectively. Developing engaging learning tools Moreover, what makes these simulations unique is the inclusion of ChatGPT-enabled dynamic moderation. Unlike fixed-path simulations, the AI chatbot can answer student players’ questions, create new scenarios, role-play stakeholders and provide analogies or explanations on demand. Individual games also adapt to student input, offer scenario-based feedback and promote strategic experimentation. ChatGPT’s feedback is personalised, aligning with the strategic logic
Ambition • ISSUE 2 • 2026 31
Made with FlippingBook - Share PDF online