How a CEO digital twin turns corporate strategy into a live, high-stakes learning experience

By Ha Hoang, Associate Dean for Research at ESSEC Business School.

What happens when students stop reading cases and start talking to them? I developed a new AI-based case experience that invites learners to do exactly that with a digital twin of a real CEO to debate the future growth strategy of his company. 

Given the early state of AI adoption by educators, the format I describe below addresses two principal concerns. On the one hand, introducing AI into the classroom risks fueling students’ over-reliance on artificial intelligence. On the other, instructors are lamenting that students are less willing to engage with lengthy cases that are central to case-based strategy instruction. The CEO agreed to participate in the development of the case because he credited his business education with his success and saw an opportunity to help keep it relevant and engaging for new students.

Growth Strategy with Moltiply CEO Digital Twin places learners in the role of strategic advisors to Moltiply Group, a publicly listed financial services company operating across multiple geographies including France, Spain and Germany. Rather than working from a traditional written case, participants interact with the digital twin of Alessandro Fracassi, co-founder and CEO of Moltiply.  With the help of Avris Technology, the twin was created and trained on publicly available financial disclosures and interviews reflecting the company’s strategic dilemmas as well as its history and origins as an Italian online comparison site founded in 2000.  The twin thus presents the CEO’s distinctive point of view and invites further inquiry so that students actively construct their understanding of the company.  

Students can engage the digital twin through voice or written dialogue, asking questions and exploring how strategic concepts are applicable to a real-world context. Given the source material, the interaction with the bot can cover a great deal of ground that allows for a variety of substantive topics to be covered including whether and where Moltiply should pursue further international expansion, growth through alliances or engage in further acquisitions.

I chose to address the issue of over-reliance on AI by linking the digital twin interaction to live, structured debate. Teams are assigned a strategic question and designated to affirm or oppose the proposed course of action. This format challenges students to examine the strategy and defend their position convincingly, building on conversation with the twin but validated by their own analyses. The overarching goal is for students to develop the ability to reason under uncertainty, engage with executive perspectives and communicate clearly and persuasively. I also ask students to vote for the team in each debate that makes the most compelling case in order to raise the stakes and engagement in class.

To facilitate debrief of the exercise, teams are asked to produce a short reflective memo documenting how they organized their inquiry with the digital twin, collaborated as a team and translated analysis into oral arguments. Most teams learned quickly that it was not intended to be an all-knowing oracle which prompted them to do further research and find additional sources of information. When a team does not sufficiently verify the digital twin’s responses, the debate format can quickly reveal this as a liability in the rebuttal or Q&A segment of the exercise. In my debrief, it can launch a discussion about the importance of critical thinking while acknowledging the benefits of AI.

Overall, I found the use of a CEO digital twin is novel and engaging for students and coupling it to the debate format allows its use to be firmly rooted in topics relevant to corporate strategy. Given its flexibility, instructors can use the case to explore other course topics, embed it within other formats including hybrid and broaden the population of learners. When linked to dissemination through the Pedagolab blog, such efforts will continue to improve our understanding of effective practices in AI-based instruction.