One approach for understanding how organizations assess technology
One of John Nash’s practical theories provides a lens for evaluating how modern organizations adopt, execute and pay for advanced technology.
Many in healthcare are familiar with the 2001 film A Beautiful Mind, in which Russell Crowe masterfully portrayed the brilliant mathematician, Professor John Nash. While the movie focused heavily on his personal struggles, Nash’s enduring legacy is his groundbreaking work on strategic decision-making and, specifically, the concept of a Nash equilibrium.
While Nash’s framework is usually confined to academic economic circles, the rapid rise of artificial intelligence in corporate environments has suddenly made Nash's theories very practical. It provides a precise lens to evaluate how modern organizations adopt, execute and pay for advanced technology.
This is playing out right now in the technology sector, highlighted by recent headlines regarding Microsoft dictating that its employees refrain from using Claude models because of the exorbitant expense of token consumption, even though those employees heavily preferred Claude over Microsoft's own Copilot.
This friction perfectly highlights the hidden game of token economics. When an organization mandates a lower-complexity, high-frequency tool like Copilot for simple inline tasks, but a worker realizes that a high-complexity agentic tool like Claude can autonomously reason through and refactor entire systems, a massive misalignment occurs.
An inline task is handled in the moment, right where your cursor is resting. Because agentic workflows require massive context windows, they consume tokens at an exponentially higher rate. If the underlying rules of the corporate game force workers into an inferior tool to save money on immediate token costs, it drives a suboptimal result for the entire enterprise.
Understanding friction points
To understand why these friction points occur, it’s crucial to look at how the individual strategies of the workplace interact.
The efficiency and productivity of AI usage within the workplace can be directly applied to a common economic concept known as game theory. Game theory involves studying strategic decision making and how rational individuals or groups of individuals maximize their own outcomes, with the final results depending on both their own choices and the choices of others.
The concept of game theory can be used to analyze a vast number of situations and scenarios that involve decision making based on what other individuals may do. Within economics, game theory can be used to study pricing decisions and negotiations. Game theory is also commonly used to study world affairs such as military strategy and diplomatic decisions.
Major developments within game theory were made by John Nash, including a theory he pioneered called a Nash equilibrium. This occurs when each participant in a certain scenario chooses their best strategy, given what the other participants decide, in a way in which no entity can improve their outcome by changing their strategy alone. In a Nash equilibrium, every player is choosing their most optimal response to the other individual, meaning no one has an incentive to switch strategies within the equilibrium.
Impact in the workplace
The concept of a Nash equilibrium can be examined within token AI usage in the workplace. To fully analyze this scenario, three different players are involved – the employee, the manager and the organization.
These players either have the option to use or encourage the use of AI, or to not use or discourage the use of AI. The outcomes of each player will be analyzed to determine the true Nash equilibrium of this scenario.
In the case of the employee, they can either use AI or not use AI for the tasks they are expected to complete. Employees are focused on completing tasks as efficiently and quickly as possible, and they are generally not concerned with the token costs that result from the use of AI. Because of this, their rational strategy is to use AI as much as possible to maximize efficiency.
In the case of managers, they are concerned with the productivity of their workers and how innovative they are. If managers encourage the use of AI for their workers, the workers become more productive, which in turn makes the manager more successful. Therefore, managers’ most rational strategy is to encourage the adoption of AI usage for their employees.
Finally, the organization is responsible for paying the bills for token AI usage, which must be factored into its rational strategy. Even though AI use by employees may create additional costs, the value of the increased productivity and innovativeness that may result from the use of AI will outweigh the costs associated with adoption.
With this, the Nash equilibrium from this scenario can now be visualized. The Nash equilibrium occurs when employees maximize the usage of AI, managers encourage and maximize the adoption of AI usage for employees, and organizations allow the adoption of AI while tolerating the additional spending that occurs.
These three decisions create a Nash equilibrium because if any of these individuals change their strategy, it will harm their outcome. If any of these players change their strategy, employees would become less productive, managers would perform worse with less productive workers, and organizations would lose overall productivity while risking slower innovative activity within the workplace.
Therefore, it’s clear why these three strategies involving the adoption of AI represent the most effective and rational decisions within the workplace.
Charting the options
Below is a three-player payoff matrix representing the scenario described above and explicitly representing the Nash equilibrium. The employee will be represented as the row player, the manager will be represented as the column player, and the organization will be represented as the outside player of the payoff matrix.
Each individual value within the payoff matrix represents the outcome, or payoff, that each player receives from making a certain decision. As can be seen in the top left box of the first payoff matrix, each player receives a payoff of 3 when the employee uses AI, the manager encourages AI use and the organization allows the use of AI.
Because every player receives their highest payoff within the matrix when these decisions are made, this outcome represents the Nash equilibrium because no player has an incentive to change their strategy independently.
Ken Poray is CEO of Integrex Health and Chair of the AI Community of Practice at the American College of Health Data Management. He has 20 years of experience working with payers, status, EDI transactions and most recently with AI workflow. Ryan Poray is a third year economics major and general business minor at the University of Maryland, and an Economics Departmental honors student.