top of page
Search

Digital Twins in the Boardroom: From "What-If" to "How".


by

Ayenda Kemp, Ph.D



The idea of using digital twins for applications is gaining serious traction (see Deloitte post). Digital twins create simulated replicas of physical system to run "what-if" scenarios. Digital twins have a growing number of operational applications, and have been used to create in silico replicas of manufacturing processes and supply chains.


But what if we could build a digital twin not just of the company's operations, but of the board itself? My recent research conversation with Dr. Karen Schnatterly dove into this complex challenge. Karen is professor at Pamplin College of Business, and an expert on board dynamics.


Our conversation explored the jump from theoretical possibility of board level digital twins to practical, insightful details needed to produce a digital twin as an organizational simulation. This post sketches the contours of that conversation.


Digital Twins and Organizational Simulation

A digital twin is a dynamic, virtual replica of a physical entity (like a factory, a city, or in this case, a board of directors) that updates with data from its real-world counterpart. This allows users to test interventions, predict outcomes, and monitor performance without impacting the real system.


In a broader sense, the board digital twin is an advanced form of organizational simulation, using sophisticated Large Language Models (LLMs) as the cognitive engines for the simulated directors (or "agents")(Doshi et al., 2025; Harrison et al., 2007).


This differs from traditional modeling in its emphasis on interaction, knowledge, and behavioral rules.Instead of focusing on broad demographic variables (like tenure or age) alone, the real power of the simulation comes from defining the agents by their knowledge base and the rules of interaction. As Ayenda Kemp explained:


"Instead of thinking of demographics, I just think knowledge set, right? So I just put R&D there. I put whatever the department would be... We're imagining that the directors have knowledge and then we're coming up with some sort of theor[y], some system of rules to think about how that knowledge is applied in the actual set sets of conversations."


Digital twins have value to the extent that the can help with organizational effectiveness. So, for a board of directors, the ultimate goal of a digital twin is to simulate and improve decision-making dynamics. As Karen Schnatterly noted:


"If boards are using digital twins to ask what-if questions about the company, why can't we use digital twins to ask what-if questions about the board's decisions about the company?"


This takes the digital twin beyond merely operational forecasting and places it squarely in the domain of organizational simulations.By connecting LLM agents to rich knowledge bases (like 10-Ks or proprietary data), they can generate realistic, knowledge-driven discussions and strategies, essentially allowing researchers to "simulate an analyst call" or a specific committee meeting for a built-in check against real-world data.



The Missing Realities: What the Conversation Revealed

While the Deloitte post focuses on the utility of the digital twin for strategic artifacts, our conversation highlighted the major, often-overlooked challenges and realities of creating a truly valuable board twin—issues that are inherently behavioral and theoretical. We highlighted three practical obstacles to operationalizing a digital twin.


1. The Conflict Between Theories of Interaction

A simulation requires making explicit assumptions about why a director speaks up, or holds back. This is where theory is critical. We noted a fundamental collision between two organizational theories (Hambrick & Mason, 1984; Kogut & Zander, 1996):


  • Knowledge-Based View (KBV): This perspective focuses on the value of sharing knowledge for mutual benefit—the risk/reward of knowledge spillover. The assumption is: "How do I lose or gain in terms of my future economic outcomes for sharing this knowledge?"

  • Upper Echelons Theory / Fault Lines: This perspective focuses on how demographic or social variables (age, gender, background) create social rules and willingness to share. As Karen noted: "like fault lines would say when you're an outgroup, you won't speak up. When you're an ingroup, you will."


The realistic digital twin must integrate both sets of rules—the logic of what a director knows, and the social logic of when and if they are willing to share it (Schnatterly et al., 2021).


2. The Necessity of Behavioral Logic and Tweakable Rules

The true benefit of a board digital twin is not just in modeling reality, but in the counterfactual—testing what would happen if you changed the behavioral rules.

"The true benefit of the digital planning thing actually is that you can... say if we were to change some of these what would have happened..." - Ayenda Kemp

This means building in logics that reflect human behavior, such as:


  • The Newcomer Effect: Are you going to "come on and wait and sit and listen" or immediately assert your expertise?

  • Tokenism: If you are the only woman on an 11-member board, does your sharing profile change compared to being one of six?

  • Interpersonal Dynamics: Simulating the impact of underlying personal conflict, where directors might consider: "What are you going to do given that the CEO is your cousin? What are you willing to do given that you hate that director over there?"


3. The Challenge of Capturing "Thought" Data


To ground the twin in genuine director expertise, we discussed the need for low-stakes, private interaction data to build an accurate "intelligence" profile. The ideal scenario: directors sharing their private, anonymized interactions.


This data, which captures how directors privately manage and process strategic situations, could be used to build the director's unique cognitive twin, moving beyond purely publicly reported data. This offers value to the boards themselves, as it allows them to better evaluate their own decision-making processes in a safe, privatized environment.



What comes next? From idea to implementation.

Our conversation made clear that the future of the board digital twin lies in the rigorous articulation and modeling of these underlying behavioral assumptions.

Karen emphasized the need for developing theory-driven rules of interaction based on existing research (e.g., on gender, fault lines, and new board member socialization). One approach is to compare ways of modeling these different theoretical logics using LLMs to see how they "collide," building the foundation for a truly powerful, insightful, and theoretically grounded digital twin of the modern corporate board.


References





 
 
 

Comments

bottom of page