When I initially set out to transform my enterprise, my strategy focused almost entirely on the individual. I wanted staff members to take complete ownership of their duties and manage themselves independently. The outcomes were decidedly mixed. With the clarity of hindsight, I realize I should have devoted far more energy to reshaping the formal and informal company policies that would have provided workers with a clearer and more visible strategic compass.

Changing The Rules Of The Game

Nick Chater, a professor of behavioral science at Warwick Business School, alongside behavioral economist George Loewenstein of Carnegie Mellon University, conceptualize this dichotomy as the difference between the i-frame and the s-frame. In a 2023 study published in Behavioral and Brain Sciences, they argue that behavioral science has concentrated too heavily on modifying individual actions (the i-frame) rather than altering the overarching environments where people function (the s-frame).

During a debate hosted by Behavioral Scientist, Chater argued that systemic architecture holds far more weight than individual agency. No matter how hard people try, they will continuously flounder within a structural framework that actively works against them.

While other academics during the debate questioned whether focusing on individuals detracts from systemic reform—arguing that both elements must be combined—my own leadership experience firmly aligns with Chater’s view that the system must take precedence.

From Experiment To Machine

Although Chater’s insights address public policy, identical challenges surface across corporate landscapes. I recently discussed the delicate balance between individuals and systems with Paddy Lambros, the founder and CEO of Dex, an AI recruitment startup launched 18 months ago. For now, Lambros has chosen to bet heavily on people: two-thirds of his personnel are former founders or hail from founding teams.

This approach stems from the necessity for rapid execution. “Businesses make a trade off between speed and predictability,” Lambros explains. “It is about being as flat and delegated as possible and requires people with high autonomy, accountability, determination and resilience.”

Instead of engineering a rigid system upfront, Lambros prioritizes organizational learning. “The machine I am building at the moment is an experiment machine,” he notes. “The rate I can ingest information, work out what is true, is the rate my company moves forward.”

Ultimately, the i-frame serves as the launchpad for innovation. Yet, once a concept proves viable, scaling it demands a dependable system. The persistent challenge lies in how that system inevitably evolves over time.

When The System Serves Itself

“There are very few examples of a larger organization or government transforming to be more agile or retaining innovative behaviors over time,” Lambros observes. “Large organizations or governments are there to sustain themselves, they trade risk for certainty, which makes them unsuitable for trying to change.”

High-growth ventures push decision-making down to the personnel executing the work. Conversely, massive enterprises and government bodies often require 15 layers of approval to confirm a single choice. While this bureaucracy creates the illusion of risk reduction, it severely caps the upside. Ironically, what leadership views as risk management frequently introduces far greater long-term danger.

A structural setup engineered strictly for certainty also tends to penalize top value creators. Price’s law, originally observed by Derek de Solla Price, indicates that the square root of a group's total headcount generates half its total output. In a 100-person team, that amounts to just 10 individuals. Furthermore, research by Rob Cross and colleagues reveals that 3% to 5% of employees drive between 20% and 35% of value-added collaborations—a pattern closely mirrored in informal network analyses conducted by my team at HumanDynamics.

While human beings remain unpredictable in isolation, they act predictably in aggregate. This means the systemic environment, rather than individual effort, dictates how the majority behaves. As Lambros asserts: “Individual action does matter, but individuals can’t be responsible for the system.”

Workers cannot rewrite rules that remain outside their control. Establishing the systemic framework is the primary responsibility of leadership, and it is a duty that cannot be delegated.

As a former CEO, I was the sole individual capable of overhauling company policies, yet I chose to pressure employees to adapt instead. Sadly, many modern AI transformation initiatives follow this exact blueprint. Demanding that employees upskill places the burden squarely on their shoulders, while the underlying processes, incentives, and decision-making rights guarded by leadership remain untouched. As Chater emphasizes, beneficiaries of the status quo naturally promote the narrative that flaws lie within individuals, effectively deflecting scrutiny away from the system itself.

Leaders committed to driving genuine behavioral change must begin by overhauling the system. For Lambros, this involves decoupling coordination from control. He implements the broadest possible boundaries, then empowers individuals to “architect and create the structure and when issues arise, let them solve them.”

Within every enterprise, a small cohort of contributors generates a disproportionate share of total value. The critical question for modern executives is simple: Does your operational system help these vital players thrive, or is it encouraging them to pack their bags?