AI INTERACTION
SERIOUS SIMULATIONS
as frameworks on innovation leadership popularised by former IBM CEO Sam Palmisano formed the basis for this simulation. It seeks to deepen students’ understanding of innovation uncertainty and path dependency, while also exploring how to balance R&D investment between current operations and future growth. Portfolio Strategy : This game stems from the Harvard Business School case study, Corning: 156 Years of Innovation , authored by Harvey Kent Bowen and Courtney Purrington. It offers students the chance to get to grips with designing and managing innovation portfolios that can account for external shocks, while also showing how portfolio diversity supports long‑term innovation resilience. Biocon – the Commercialisation Gamble : a simulation inspired by Harvard Business School professor Gary Pisano’s paper, Science business . Students learn how to manage risk in the biotech industry and choose the right commercialisation option, navigating trade-offs around licensing and product market entry.
A brief introduction to five custom ChatGPT simulations created for a course on strategic innovation, as well as the research frameworks on which they are based . Survive the Dominant Design : A game highlighting that while an early market entry can be advantageous, long-term survival rests on how well a company adapts its strategy during the exploratory pre-dominant design phase. It is inspired by the research paper, Innovation, competition and industry structure , by MIT Sloan School of Management’s James Utterback and Fernando Suárez. Tushman (the Ambidextrous CEO) : Taking Michael Tushman and Charles O’Reilly’s paper on ambidextrous organisations as its cue, this simulation shows how competence-destroying discontinuities threaten firm survival. It also helps students learn to balance exploration and exploitation through structural or contextual ambidexterity. Exploration vs Exploitation : The myopia of learning by Wharton’s Daniel Levinthal and Stanford’s James March, as well
behind each student’s decisions and helping learners articulate their reasoning. In this way, games function as a co-instructor, capable of adjusting to each student’s level of understanding and employing Socratic questioning to encourage deeper learning. The custom GPT is prompted to behave rationally and provided with complete information, including when to deploy ‘shocks’ during the process. These shocks are unannounced disruptions, such as competitive moves, market shifts or internal crises, that force students to reconsider their strategies. Unlike the GPT, students play with imperfect information, making it imperative that a shock is not revealed to them ahead of time. The asymmetry in information echoes real-world decision environments, deepening the exercise’s relevance. For instance, students are tasked with rebalancing investment priorities after unexpected financial shocks in the simulation, Portfolio Strategy . The game gives first-hand insight into how difficult it is to pursue innovation without compromising operational stability and why a portfolio of innovations can help absorb shocks. As mentioned, no coding was required to create these simulations. However, their development still involved a considerable amount of work, with the greatest amount of time spent on testing. To begin with, a clear set of prompts and input instructions for the custom GPT had to be set up. With each testing round, details of what the simulation should do and which frameworks and research it should follow were then added and updated. For example, the Survive the Dominant Design simulation was prompted to play out over eight rounds, with each round representing a year. The
Ambition • ISSUE 2 • 2026 33
Made with FlippingBook - Share PDF online