^{Photo Credit: Diane Bondareff Photography}

I was nervous walking in.

I had been invited to speak at a board and CEO forum on AI, and everything on the agenda before my session was about leading at AI speed. What boards must align on now. Do we buy, build, or license AI capabilities? How AI is changing decision-making, accountability, and execution speed. How to escape pilot purgatory. Which AI bets deserve more investment and which deserve sharper scrutiny. What the EU AI Act means for U.S. companies when its obligations reach them through vendor supply chains. Sovereign versus enterprise-controlled agents. Agentic AI in the boardroom. The AI-accelerated threat landscape in cybersecurity. Using AI to simulate and stress test high-stakes decisions.

These are all genuinely important board conversations. I sat there listening and thinking, this room is moving fast, and I am about to stand up and talk about human intelligence. Would they be interested in what I had to say about the human advantage? Or would it sound like the soft session, the one you sit through politely before the next panel on capital allocation?

What happened taught me something I did not expect, and I think it should shape how boards approach AI governance: the human conversation was not the soft session. It was the one the room turned out to be equally hungry for. But first, let me walk you through what I presented, because the argument is the setup for the lesson.

Subscribe now

Are we missing something here?

I opened by asking the room what they were walking in with. What are your biggest ambitions for human capital in an age of AI? What risks concern you? What questions are you already asking, and what knowledge are you bringing?

The answers came fast, and they were good. One director who talks to hundreds of C-suite executives every month said the big theme she keeps hearing is companies stuck in replacing humans with AI rather than augmenting them. After engaging with her a bit, it had me realize that if everybody has access to the same AI, the competitive advantage you get from pure replacement is zero. Isn’t strategy supposed to create an advantage? If AI becomes table stakes, and your entire play is substitution, where does your edge come from?

Another raised retention and fear. If your workforce is riding the subway home reading a headline about their CEO considering AI replacement, what happens to their productivity, creativity, and loyalty? And we know from decades of research what fear does to a brain. A brain in fear stops learning, creating, and collaborating. If we are putting our workforce into a state of fear, what are we doing to our capacity to reimagine where we can go?

A third raised the autopilot problem: people overusing AI without using it well, putting their brains on cruise control. Another described the governance gap in tool adoption: everyone experimenting, a few champions emerging, and nobody harvesting what the champions learned to lift the rest of the organization. That was putting it nice. I think she actually used the words “total chaos”.

One CEO story captured the moment perfectly. His company had rolled out AI-driven proactive maintenance for their vehicle fleet. Sounds great. Then his entire fleet got grounded one day because there was a new “windshield wiper fluid” preventative maintenance program in which all were non-compliant! He said, I don’t care about windshield wiper fluid! I want my trucks on the job. Somebody had spent a long time building that. Who decided this was what we were doing?”

The typical categories of AI human capital governance, and there are many good ones, are primarily oriented toward workforce exposure. Either how do we protect people from irresponsible AI, or how do we prepare people to use AI productively? For sure, both are necessary. But we are much weaker on a third question:

How does the organization preserve and regenerate the human intelligence on which future performance depends?

What happens if we get this wrong? The risks compound, and I have come to think of them as six D’s. Departure, as your best people leave. Dumbing, as capability gets hollowed out. Distraction, as activity outruns strategic direction. Despair, as engagement falls. Devaluing, of both the humans and the enterprise, because if you are doing exactly what everyone else is doing, both are worth less. And Disconnection, as effort loses context and coordination.

“That Schedule Won’t Work”

Let me give you a simple example from a North American restaurant chain with over 100 locations.

Restaurant scheduling is a perfect AI use case. It is repeatable, data-rich, and time-consuming. An experienced general manager needs about four hours a week to build a good schedule. A junior manager needs much longer. So they deployed AI, and the AI built a schedule in seconds. It forecasted demand, hit labor targets, adhered to availability rules, covered every role, and optimized overtime. It had learned from thousands of schedules. Honestly, it was pretty great.

Then the experienced GM looked at it and said, “That won’t work.”

When asked why it wouldn’t work, she said, “Those two just broke up, and you cannot put them on the same shift right now. That one’s parenting schedule just changed and she has to pick her kid up at three. The shift leader you slotted has worked three closes in a row and is about to burn out. And that one is trying to get promoted, still has not run a dinner rush, and needs to be on it, because he is getting antsy and easier for another restaurant to poach if he does not feel like he is progressing.”

The schedule took seconds. That sentence took years.

The AI had studied thousands of schedules. The GM had also created thousands of schedules. But the difference is that she was accountable for the near-term and long-term results of every one of them. She lived the schedules she created. She failed, she succeeded, and she lived with the consequences, which is precisely why she was able to learn. Philosophers and knowledge researchers call this tacit knowledge, and it is a fundamentally different kind of knowing than the explicit, encoded knowledge a machine holds [1]. Michael Polanyi put it simply: we know more than we can tell.

Sure, you can keep programming in the exceptions, but don’t lose the forest for the trees on this one. The GM worked with the AI to make a great schedule and did shave off about 50% of the time. With practice, she will get more efficient.

However, there is also a Paul Harvey rest-of-the-story on this – with this AI deployment, any deviation from the AI schedule was flagged as noncompliant, and the manager had to explain it. Not to “improve the model”, but rather to ensure compliance and consistency across the organization.

Pause on that design choice. What do you do with the explained deviation? Does it become an opportunity for the human intelligence to correct and teach the system? Or does it become friction the manager learns to avoid? This GM made it work, but what about the next one? Where are we automating work or discouraging judgment that also develops the sensibilities our future leaders will need?

If you read Part 4 of this series, you already have the language for what went wrong here. That deployment applied a fixed constraint, zero deviation with a compliance flag, to a situation that called for a governing constraint, expert discretion inside defined bounds. The AI did not fail. The constraint form was misassigned. And notice what gets built while nobody is deciding anything. No policy tells that GM to stop using her judgment, but the flag teaches her to. That is a dark constraint forming in real time, and AI deployments manufacture them faster than governance can catch them [2].

Not Because I’m Nice

I want to be clear about why I care about this. I’m not a “save the humans because I’m a human and we should” kind of gal. I am a “create-long-term-value-through-a-sound-strategy-that-leverages-your-assets-and-mobilizes-your-workforce” kind of gal. And a first principle of optimization is that to optimize any system, you have to understand it.

I learned that as a brand new lieutenant in aircraft maintenance when I was handed a fleet of C-130s I barely understood. My chief would prep me before briefings: when the colonel asks this, say this. It worked for a while. He was like my early supplemental intelligence. But until I opened up the engine and actually learned the aircraft, I could not ask the right questions. Right now, most organizations are in the scripted-answers phase with their people. We are racing to understand artificial intelligence while we still do not really understand human intelligence. This is our opportunity to do both at once.

So put the two intelligences side by side.

This is a generalization, and the boundaries blur at the edges. But the point stands. These are different kinds of intelligence, and if we rank one above the other and manage the “lesser” one down, we forfeit the advantage that only comes from orchestrating both. For a century we have been trained to minimize the subjective, to organize the human messiness out of our systems. Humans are contextual, historical, relational, and gloriously weird in their interpretations. That weirdness is where innovation comes from.

Remove the Friction, Atrophy Happens

Human beings run on an energy gradient. Given two paths to the same result, we take the one requiring less effort. That is our design, and it is what makes us efficient. AI, meanwhile, is engineered to reduce friction. Frictionlessness is the product. So of course humans will use it. Telling people to preserve their thinking sovereignty through willpower alone is a losing strategy, because we are famously bad at willpower.

The problem is that human capability develops and maintains itself through effortful practice, feedback, and consequence. Remove the friction, and atrophy happens. Researchers call it cognitive offloading, and the “dumbing” cascade is worth understanding. Offloading reduces practiced capability. Reduced practice weakens skill formation, learning, and tacit judgment. Weaker judgment increases dependence on AI’s framing. Shared dependence narrows the range of perspectives, and that convergence reduces the very variance from which innovation emerges [3]. Reduced authorship gives way to diffusion of responsibility, and studies from military cadets to healthcare workers show that when people feel less accountable for a decision, less ethical behavior gets easier [4]. Diffused accountability then makes further offloading easier still. The loop feeds itself.

This is the argument I made in Part 4 of this series, wearing different clothes. There I argued that constraint is generative, that freedom comes from the right kind of constraint rather than from its absence. Friction is constraint operating at the level of individual cognition. Strip all of it out and you do not get a faster thinker. You get an atrophied one, for the same reason that removing every constraint from your diet does not produce health.

This is not a theoretical cascade. The evidence is arriving now rather than on the futurist timeline. Read it carefully, because these are different designs measuring different things, and none of them follows a workforce for a decade. Taken together they still point one way:

• In a study at the Royal Military Academy, cadets acting as drone operators had their moral decisions measurably shifted by aggressive and conservative AI recommendations, and they reported feeling less personally responsible for outcomes even though the final decision technically remained theirs [5].

• A 2025 paper in Nature covering thirteen experiments found that delegating tasks to machines increased unethical behavior, particularly when people could induce wrongdoing through higher-level goals without explicitly instructing the machine to act dishonestly [6].

• Gerlich’s 2025 study of 666 participants found a strong negative correlation (r = -0.75) between AI tool usage and critical-thinking scores, mediated by cognitive offloading [7]. That study is cross-sectional and self-reported, so it establishes an association, not a cause.

• Nataliya Kosmyna’s 2025 EEG work at MIT found that writing with LLM assistance produced measurably reduced neural connectivity in the brain regions responsible for coherent argumentation and sustained attention. That one is a preprint with 54 participants and a published methodological critique, so I hold it more loosely than the others, yet still worth paying attention to [8].

• And Cheng et al., published in Science in 2026 across three preregistered experiments with 2,405 participants, demonstrated that a single interaction with a sycophantic AI model reduced willingness to repair interpersonal conflict and increased users’ conviction that they were right [9].

• In the same study, across eleven frontier models, AI affirmed users roughly 50 percent more often than humans did, including on queries involving deception, illegality, or other harms.

While these are different studies, with different methods, across different timeframes, they point in the same direction: capability erosion as apprenticeship and judgment development are bypassed; epistemic capture as thinking sovereignty narrows and perspectives converge; and enterprise consequence as innovation portfolios compress while accountability diffuses.

We have known a version of this for decades. It is called the retirement effect: when cognitive stimulation drops, cognitive function declines. Work is full of problem solving, decision making, and learning, and our brains need that engagement to stay healthy [10]. Use it or lose it was never just about muscles.

I will confess we lived this in my own company. An unfortunate event cost us some of our human capability right as AI tools were maturing, and we had to serve the same clients at the same level with half the people. AI made that possible, and we got really good at it. Then I started noticing the product. It was not connecting the way it used to. So we changed our practice. We now do the thinking first, then go to AI, then come back and improve our own work. We think on paper first now, and we learn AI at the same time.

It is important that we ask what our current AI portfolio is actually designed to optimize, and for whom. Faster answers? Better answers? Better sensing? Better decisions? I come from the military, where speed was always precious, but a faster answer never automatically meant a better one.

Humans Care. And That Is the Advantage.

Humans care. AI does not. And by care I mean something more than affection or affinity, though we use the word that way in everyday speech. The philosopher B. Scot Rousse describes it this way:

“Care is more than a feeling of affection. It is a way of being involved and meaningfully engaged in the world.” [11]

Care is what directs attention. It is where discretionary effort, commitment, prioritization, and responsibility come from. If you think about it, “care” just may be the fundamental structure of human being; nothing shows up as mattering to us at all except through it. Care cannot be installed in a model. It can only be cultivated, or eroded, in people. Which means that if we disengage our people, the care goes down, and with it goes the workforce capability we were counting on.

After twenty-five years of being called in after the reorg, after the acquisition, after the transformation stalled, I can tell you where change efforts reliably break, every time. They rarely fail on strategy or technology. They fail because leaders never addressed the human cares. The track record is sobering. In McKinsey’s own survey work, fewer than a third of companies say their transformations both improved performance and sustained the improvement, and what separates the successes is not which actions they chose but how many of them they finished. The acquisition record is no kinder, with most deals failing to lift shareholder returns and key talent leaving inside the first year [12]. None of that is an AI problem. It is a scaled-change problem. So why would AI transformation, the largest change effort most of these companies will ever attempt, turn out differently?

William Bridges made a distinction leaders keep relearning the hard way: managing a change is different work from transitioning the humans through it [13]. Change management handles the ordered side, the technology, workflow, policies, platforms, and business case. Transition is what happens inside people, and it runs on different questions. Will I still be good at my job? Who am I if I am no longer the expert? Is there a home for me here? What do humans care about? Contributing. Their future. Their identity. Belonging. In our experience, people stop engaging and eventually leave for one reason at the root: they can no longer see a future for themselves where they belong, with an identity they are proud of, and clarity on how their contribution connects to results.

That is why AI cannot be treated as only a change to manage. For many of your people, it is an identity transition. Many of us built our identities out of expertise that AI is now taking on. Another question: what future might your workforce be inferring from your AI investments, whether or not you have declared one?

One of my favorite movies is Hidden Figures, and one scene has stayed with me for years. Dorothy Vaughan, watching the IBM mainframe arrive that would replace the human computers, does not wait to be replaced. She teaches herself FORTRAN, then teaches her whole team, and re-places herself and her people inside the new future. That distinction, replaced versus re-placed, is the leadership opportunity in front of every executive team right now: to give your human capital a new vision in which their judgment, care, and courage matter more, not less.

An Intelligence Strategy™, Not an AI Strategy

So what do you do with all this at the governance level? Here is the reframe we are working with. Rather than just an AI strategy, build an intelligence strategy.

An intelligence strategy protects and grows the company’s capacity to create value by orchestrating human and artificial intelligence in concert. Artificial intelligence brings scalable retrieval, analysis, prediction, pattern detection, simulation, and generation. Human intelligence brings situated experience, contextual judgment, relationships, imagination, meaning-making, and accountability. The strategy is the design of how each amplifies the other.

With an intelligence strategy, the executive question shifts from “Where can we deploy AI?” to “Where could better intelligence materially change outcomes and enterprise value, and how should human and artificial intelligence work together to produce it?”

For boards and executive teams, we frame the work in five connected areas surrounding one center: your enterprise intelligence advantage.

• Strategic ambition and advantage. Where will human and artificial intelligence materially improve our competitive position, and are our investments coherent with that ambition? What are we actually willing to bet about AI opportunity, risk, and future value creation?

• Intelligence assets and architecture. Which of our intelligence assets are distinctive, meaning not easily bought, copied, or generated? How are we protecting them from loss, commoditization, or platform dependence? What should machines retrieve, analyze, predict, and generate, and what must humans sense, interpret, judge, and own?

• Capability and workforce. Where could automation weaken the judgment, expertise, succession, or accountability the enterprise needs? How will we redesign work so human judgment and leadership continue to grow? And what credible future can people see themselves contributing to and belonging in?

• Governance and accountability. What will we not delegate? Where must an accountable human always remain? Who owns the design and performance of the intelligence system and each consequential AI-informed decision? What evidence will show that decision quality, capability, and enterprise value are actually improving?

Do you sit down and answer every question in one session? Maybe not. The point is to have these conversations with your executive team and let the answers shape your strategic bets. One panelist that same day observed that the AI pilots that succeed are the ones with a clear human owner. That is the frame in miniature.

What the Room Taught Me

Now back to my nervousness.

The talk was received more warmly than I could have hoped. But the moment that stopped me came near the end, when a director named Michelle reflected on the day. She said something like: when I talk to employees about having been here, I am not going to say I was at an AI conference learning about the machines. I was at an intelligence conference. That one sentence reframed the entire agenda. Buy, build, or license. Governance. Cybersecurity. Workforce redesign. All of it is intelligence strategy, intelligence implementation, intelligence governance. AI simply shows up differently when you hear it inside that context.

What really struck me about the day was that Board Directors care in the way that humans care. Underneath the fiduciary language and the capital questions, the people in that room were asking about their own kids choosing what to study, about the people in their companies riding the subway home wondering whether to look for something else, about their own identities and futures. I walked in worried that the human conversation would be unwelcome at AI speed. I walked out understanding that it was the conversation the room was hungry for, because the people governing our largest enterprises are humans first, and they know what is at stake.

Why talk about human intelligence at an AI forum? Because when I asked myself what I am really standing for, the answer was our human future. Governance is not only about preventing AI from doing harm. It is about creating the conditions in which AI makes the enterprise more intelligent and people can see a consequential place for themselves within it. That future will have to be designed. I left the forum believing the people in that room are capable of designing it, and I am hopeful we will.

Next in the series: how surveys conceal lived truth, and why the gap between reported experience and lived experience contains your most valuable intelligence.

What future might your workforce be inferring from your AI investments? Hit reply and share your thoughts. I read every response and often explore these insights in future articles.

References

[1] Polanyi, M. (1966). The Tacit Dimension. University of Chicago Press.

[2] Snowden, D. J. (2016). Freedom through constraints, and Dark constraints. The Cynefin Company blog. Constraint typology at cynefin.io/wiki/Constraint_typology. The underlying concept of constraints in Cynefin draws on Juarrero, A. (1999). Dynamics in Action. MIT Press.

[3] Risko, E. F., & Gilbert, S. J. (2016). Cognitive Offloading. Trends in Cognitive Sciences, 20(9), 676-688.

[4] Bandura, A. (1999). Moral Disengagement in the Perpetration of Inhumanities. Personality and Social Psychology Review, 3(3), 193-209.

[5] Salatino, A., Prével, A., Caspar, E. A., & Lo Bue, S. (2025). Influence of AI behavior on human moral decisions, agency, and responsibility. Scientific Reports, 15, 95587.

[6] Köbis, N., Rahwan, Z., Rilla, R., Supriyatno, B. I., Bersch, C., Ajaj, T., Bonnefon, J.-F., & Rahwan, I. (2025). Delegation to artificial intelligence can increase dishonest behaviour. Nature, 646, 126-134. https://doi.org/10.1038/s41586-025-09505-x

[7] Gerlich, M. (2025). AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking. Societies, 15(1), 6. https://doi.org/10.3390/soc15010006 Corrected 10 September 2025; figures cited here are taken from the corrected version.

[8] Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X.-H., Beresnitzky, A. V., Braunstein, I., & Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task. arXiv:2506.08872. Preprint, not peer reviewed. See also Stanković, M., Hirche, E., Kollatzsch, S., & Doetsch, J. N. (2026). Comment on: Your Brain on ChatGPT. arXiv:2601.00856.

[9] Cheng, M., Lee, C., Khadpe, P., Yu, S., Han, D., & Jurafsky, D. (2026). Sycophantic AI decreases prosocial intentions and promotes dependence. Science, 391(6792). https://doi.org/10.1126/science.aec8352

[10] Rohwedder, S., & Willis, R. J. (2010). Mental Retirement. Journal of Economic Perspectives, 24(1), 119-138.

[11] Rousse, B. S. Without Why (Substack). https://withoutwhy.substack.com/

[12] McKinsey & Company (2021). Losing from Day One: Why Even Successful Transformations Fall Short. Note that the widely repeated “70 percent of transformations fail” figure has no documented empirical basis; see Hughes, M. (2011). Do 70 Per Cent of All Organizational Change Initiatives Really Fail? Journal of Change Management, 11(4), 451-464. On acquisitions, see Christensen, C. M., Alton, R., Rising, C., & Waldeck, A. (2011). The Big Idea: The New M&A Playbook. Harvard Business Review.

[13] Bridges, W. (2004). Transitions: Making Sense of Life’s Changes. Da Capo Press; Bridges, W. (2009). Managing Transitions: Making the Most of Change. Da Capo Press.

That is how I describe the beginning of my own transition in the preface of my new book out today: Re-Placed: Answering the Call of Leadership in the Age of AI.

While I may have called it by different names, my focus has always been how to unlock the intelligence and agency that already existed in people and organizations, but was trapped inside legacy systems. But now, a bit of a curve ball with this new “intelligence” has arrived.

We are now dealing with an existential technology. Nonhuman intelligence is participating in our work and our decisions, but it does not participate in our humanity. It does not care, suffer, belong, identify, or take responsibility. This is not just another technological adoption. It is not just a call to reskill. It forces us to question our very “place” in the world.

I wrote Re-Placed because I saw leaders getting caught in a new tension - they are excited about the capability of AI, but they are also dealing with “ontological vulnerability”—the experience of having the foundations of your identity, your value, and your place in the world suddenly become uncertain. When we talk about being “replaced” by AI, we are not talking about substituting an algorithm for a human worker. We are altering our foundational standing.

But the threat of replacement opens the possibility of re-placement. It is the chance to consciously claim a new position from which our distinctly human capacities matter more, not less. Being re-placed is the antithesis of being replaced.

Re-Placed is a practical orientation for leaders. It provides a transition framework and a pathway to reinvention so that you can lead your people into a new era, rather than just coping with the end of the old one.

If you carry the responsibility for guiding an organization through what comes next, I wrote this for you.

You can order your copy of Re-Placed here.

I am also conducting ongoing research into how leaders are experiencing this transition on the ground. I want to hear what you are actually dealing with. Please take a few minutes to participate in the research survey: https://karizeller.com/survey

Thank you,

Kari

Earlier in the series:

· Part 1 – “Trapped Genius: Why Smart Organizations Can’t Access Their Smartest People”

· Part 2 – “Your Organization’s Best Ideas Are Stuck in the Middle”

· Part 3 – “From Bottlenecks to Hubs: Why Your Middle Managers Are Your Secret Intelligence Weapon”

For over 700 years, the sonnet has been a crucible for literary genius. A rigid form of fourteen lines, a strict meter, and an intricate rhyme scheme; it should, by all accounts, be a cage. Yet for masters from Petrarch to Shakespeare, it became a powerful tool. The constraints of the form did not suppress their brilliance; they channeled it, forcing a discipline and intensity that produced some of the most enduring works in Western literature.

This same principle applies to the Gothic cathedrals of Europe. The builders of Chartres and Notre Dame were bound by the brutal constraints of stone, gravity, and medieval engineering. From these limitations, they innovated the ribbed vault, the pointed arch, and the flying buttress—architectural breakthroughs that allowed them to build structures of breathtaking light and height.

Every organization lives inside a constraint ecosystem. Some constraints are formal: budgets, approval thresholds, reporting lines, and compliance requirements. Some are temporal: quarterly cycles, product launch calendars, promotion windows. Some are cultural: what is praised, what is remembered, what quietly ends careers. Some are narrative: who we believe we are in the market and what kind of risks “people like us” take.

We rarely see these as a system. They feel like background conditions; yet they determine what is possible, what is probable, and what will never be attempted. The constraint ecosystem shapes the intelligence of the enterprise.

Without constraint, nothing progresses.

That may sound counterintuitive in an era that prizes empowerment and agility. But consider something simple. If you remove all constraints from your diet, you do not create health. If you remove all timing constraints from a workout regimen, you do not create athletic performance. Even something as basic as intellectual work depends on constraint. For example, a deadline focuses thinking or a budget forces prioritization. A uniform and strict schedule, as I learned when I was a cadet at the Air Force Academy, can free cognitive bandwidth. When I did not have to decide what to wear or where to be, I was surprised at how easily my brain could think about other domains.

I learned that freedom comes from the right kind of constraint, not the absence of it.

An organization without intelligent constraints is a place where brilliant people are set adrift. When every path is possible, no path has meaning. When every project is a priority, nothing is. When this design failure occurs, we think our top talent is burning out from overwork, but rather, they are experiencing dissipating energy and fragmented genius into a thousand shallow initiatives.

Different Situations call for Different Constraint

Within any serious enterprise, very different kinds of situations coexist.

A hospital is an obvious example. In the operating room, there is no room for interpretation of the sterilization protocol. Hands are washed in a specific way; instruments are handled in a defined sequence; checklists must be followed. In the Cynefin Modelⁱ, these are all called fixed constraints and the purpose is reliability and consistency. The value comes from eliminating variation because deviation is known to introduce unacceptable risk. I like to think of the Grand Canyon – the water is only going in one direction because the constraints of the canyon walls are unforgiving.

In the same hospital, interpreting a complex scan requires expertise. There are diagnostic guidelines and escalation protocols, but no single mechanical answer. Radiologists operate within governing constraints. Clinical standards define acceptable bounds, decision rights clarify who makes the call, different experts are called in. Within those parameters, subject matter expertise, regulations, and judgment come together to create disciplined discretion rather than uniformity.

Then there is the hospital’s long-term positioning. How it responds to shifting reimbursement models, how it integrates telemedicine into its operating model, how it competes for specialist talent. Here, no amount of analysis can fully predict outcomes. Leadership may articulate strategic trade-offs, operating principles, or what the military calls commander’s intent: what must be achieved, what must be protected, what risks are acceptable. These are enabling constraints. They do not specify procedure. They shape direction and allow solutions to emerge through interaction. There is no right answer or even a known answer. Constraints in this situation are more principle-based than control-based.

Fixed and governing constraints are best for the knowable, ordered scenarios, whereas enabling constraints are best for the unknown, complex scenarios. Like most organizations, the hospital not one or the other, but rather it moves constantly between situations that demand different forms of discipline.

In fact, nearly all organizations contain all three realities at once.

In payroll processing, repeatability is the objective. In supply chain optimization, expert modeling matters. In redefining customer experience in response to new digital behaviors, the solution must be discovered through iteration.

The mistake is not really too much or too little constraint, but rather, failing to intentionally design constraints in relation to the nature of the situation. Or what I often see in client situations is an organization that has outgrown an outdated constraint framework.

Leaders often assume discipline has one form. Either tighten control or loosen it. But discipline in a routine compliance process does not look like discipline in a cross-functional transformation. In one case, discipline means eliminating variation. In the other, it means structuring variation so learning can occur.

When leaders attempt to manage a strategic repositioning the same way they manage expense approvals, the effect is subtle at first. It feels right, but the results are predictable and disappointing:

• Milestones are defined before the problem has taken shape.

• Metrics are locked in before meaningful patterns have emerged.

• Reporting templates are standardized for initiatives that are still exploratory.

Teams respond rationally:

• They narrow proposals to what fits pre-existing categories.

• They avoid ideas that cannot survive a rigid stage-gate sequence.

• Exploration becomes an incremental improvement of what is already known.

Nothing appears to be chaotic. In fact, everything looks controlled. Yet over time, the organization becomes cautious in precisely the areas where adaptability is needed.

At the other extreme, leaders sometimes remove constraints in the name of empowerment. In this case, decision rights are pushed out or down, approval layers are reduced, and teams are told to innovate freely!

In this case, what is often missing is a shared directional constraint that binds those innovations together. Without clearly articulated trade-offs, operating principles, or non-negotiable outcomes, each unit quite reasonably optimizes for its own objectives:

• Marketing pursues engagement.

• Operations pursues efficiency.

• Product pursues feature expansion.

• Finance protects margins.

In other words, each group acts intelligently within its local frame. Yet without a shared attractor, those efforts do not accumulate into coherent advantage for the organization. Yes, the organization moves forward, but it does not compound. 1+1 = 2, not 5, 7 or 10.

Bottom line: constraint is not the enemy of freedom. Rather, a set of designed constraints creates a condition for coordinated action. Every organization already lives inside a constraint ecosystem. Our work as leaders is to design it deliberately rather than manage around what was inherited unconsciously.

The Constraints You Do Not See: The Narrative Field

Formal constraints are visible. They live in policy manuals, governance charts, capital allocation models, and reporting dashboards. Leaders debate them explicitly.

The most powerful constraints are often invisible.

Every organization carries a narrative field — a shared set of stories about who “we” are, how success is achieved here, and what kinds of risks are acceptable. These stories do not function as formal rules. They function as orientation. They shape what is thinkable, what is sayable, and what is strategically imaginable long before any formal decision is made.

This is what Prof. Dave Snowden, the developer of the Cynefin model, refers to as dark constraints: constraints that exert force without being formally articulated.

In some cases, dark constraints stabilize performance in ways that extend well beyond routine operations. A company may carry a deeply embedded narrative that “we are engineers first.” That identity does more than reinforce rigor in compliance processes; it also shapes trade-offs in ambiguous situations. When speed conflicts with quality, the answer is clear without escalation. When an opportunity in the market tempts overextension, the bias toward technical integrity prevails. The narrative field becomes an enabling constraint. It channels decision-making across levels without requiring constant intervention from the top. In this case, identity can generate coherence.

In other cases, however, the same mechanism limits range.

Consider an organization that has long known itself as relationship-driven and consensus-based. Decisions are expected to be socially integrated before they are formally made. The narrative field values harmony and inclusion. When the company scales or enters more competitive markets, that same identity can begin to resist sharper trade-offs. Direct accountability feels abrasive. Clear performance differentiation feels culturally dissonant. Leaders may speak openly about becoming more performance-oriented, yet the underlying story of “who we are” constrains what forms of discipline are tolerated. We see this story play out time and again when a local company is acquired by a national brand. The deeply held values are different, even if the ones on the wall are similar.

The constraint is not written anywhere. It operates through recognition and reward. Those who behave consistently with the narrative advance. Those who challenge it often experience subtle friction or exclusion.

Dark constraints also shape decision flow in less visible ways.

In one global services firm, there was no policy requiring CFO endorsement before new initiatives were presented to the executive committee. Yet over time, everyone understood that proposals without financial pre-alignment were unlikely to survive. Business leaders began withholding early-stage ideas until projected ROI could be confidently modeled. What appeared to be financial rigor slowly compressed the organization’s exploratory capacity. Ideas that required iterative discovery rarely entered formal conversation.

Leadership believed it had empowered bold thinking. The lived constraint suggested otherwise.

These tensions often remain latent until the organization attempts to shift direction. When formal strategy evolves but the narrative field does not, these dark constraints constrain progress.

An executive team may declare a commitment to experimentation and adaptive learning. Town halls celebrate risk-taking. Strategic documents emphasize innovation. Yet promotion patterns reveal that visible failure lingers in performance evaluations for years. The organization carries a long memory for mistakes and a short memory for responsible risk.

No policy prohibits experimentation. Yet the narrative field communicates caution. Execution falters not because strategy lacks clarity, but because the constraint ecosystem—formal and informal—sends competing signals about what survival and success actually require.

Dark constraints are not inherently problematic. They often encode the history of what has worked. They protect coherence. They reduce the need for constant oversight. The difficulty arises when they remain unexamined while the environment shifts. What once stabilized performance begins to narrow possibility.

Most transformation efforts fail because they adjust the formal layer and ignore the narrative field beneath it. To design constraint intelligently, leaders must account not only for governance structures and decision rights, but for the narrative fields within which their organizations operate.

The constraint ecosystem includes both.

Reading and Reshaping the Constraint Ecosystem

If every organization already lives inside a constraint ecosystem, the work of leadership is not to add more control or remove it indiscriminately. It is to read the system accurately and then adjust the form and coupling of constraint in relation to the situation.

Constraint design operates across two dimensions.

The first is structural: policies, governance, decision rights, capital thresholds, operating models. This includes fixed, governing, and enabling constraints as formal design choices — vision, strategy, trade-offs, operating principles, commander’s intent.

The second is narrative: the stories, memories, identity commitments, and promotion patterns that shape how those formal constraints are interpreted and lived.

Most transformation efforts adjust the structural layer while leaving the narrative layer untouched. When the two are misaligned, the narrative layer usually wins.

Below is a working lens rather than a rigid taxonomy.

The misapplication risk for enabling constraints deserves clarity.

When leaders articulate aspiration without specifying trade-offs, the constraint lacks operational force. “Customer first” competes with “margin discipline.” “Move fast” competes with “avoid visible failure.” In the absence of explicit hierarchy among values, local leaders resolve tensions in inconsistent ways. The constraint exists rhetorically, but not structurally. Energy disperses because decisions are not anchored to shared, lived trade-offs.

Dark constraints operate differently. They are often inherited, but they are not immutable.

Narrative fields are reinforced or reshaped through attention, language, and reward. When leaders articulate what has historically guided the organization—naming it clearly—they bring the constraint into view. When they consistently reward behaviors aligned with an evolving identity, they empower a new narrative field to take root.

Dark constraints grow where attention flows.

If an organization carries the narrative that “we protect quarterly earnings above all else,” leaders who repeatedly signal that long-term capability building matters more than short-term optics begin to alter that field—but only if incentives and promotions follow the rhetoric. Narrative fields do not shift through declaration. They shift through reinforcement.

A simple example was an encounter I had through my consulting work with a three-star general. At the time, a cultural narrative had begun to take hold that it wasn’t safe to raise perceived breakdowns. Determined to shift this, the general made a deliberate move. During one meeting, a young colonel respectfully challenged a stated direction. Recognizing how critical visible support was in that moment, the general stood up, shook his hand in front of the group, and said, “Thank you — we need more of that.”

That single act of reinforcement helped reshape how openness to feedback was interpreted. Over time, it contributed to a new narrative — one that encouraged people to surface challenges, share bad news, and speak up about breakdowns without hesitation.

Reshaping Your Constraint Ecosystem

When leaders say execution is slow, the instinct is to demand urgency. Yet slow execution frequently signals constraint friction. Controls layered over time create approval cycles and compliance rituals disconnected from actual risk. Workarounds proliferate. Energy is spent navigating the system rather than advancing strategy. Constraint language reframes these frustrations. Instead of questioning motivation, leaders can examine design.

Reshaping constraints is not a matter of issuing new policies. It is a disciplined adjustment of both structural and narrative layers.

This can be visualized as a deliberate sequence rather than a reactive response.

1. Clarify the Nature of the Situation

Is the objective repeatability, optimization within known bounds, or exploration under uncertainty?

If leaders cannot answer this clearly, constraint design will default to habit.

2. Examine the Existing Constraint Form

• Are fixed controls being used where experimentation is required?

• Are governing parameters clear enough to prevent constant escalation?

• Are enabling constraints sufficiently specific to guide real trade-offs?

• Is the narrative field reinforcing or undermining these structural choices?

3. Surface Narrative Constraints Explicitly

• What identity stories stabilize current behavior?

• What behaviors consistently lead to advancement?

• What topics generate discomfort or avoidance?

• Where does lived practice contradict declared strategy?

Leaders often discover that the formal constraint they believed was shaping behavior is secondary to the narrative field. A stated commitment to experimentation will not overcome a promotion system that penalizes visible failure. A declared strategy of integration will not override a deeply embedded identity of divisional autonomy.

4. Adjust the Coupling Across Layers

Adjustment does not mean relaxing discipline. It means aligning discipline with the situation and ensuring the narrative layer reinforces, rather than contradicts, structural design.

The Question Leaders Must Confront

Every executive team operates inside a constraint ecosystem.

Some of it has been deliberately designed—through strategy, governance, decision rights, capital filters. Much of it has accumulated over time through inherited processes, promotion patterns, identity commitments, and stories about what succeeds here and what does not.

Together, those structural and narrative constraints shape what your organization can see, how it makes sense of what it can see, what it will attempt, and what it will avoid.

Leadership at scale is less about declaring direction and more about shaping the conditions under which direction becomes real. If initiatives feel slow, if innovation remains incremental, if alignment requires repeated escalation, the issue may not lie in talent or effort. It may lie in the form of discipline that has been normalized, and the constraint ecosystem your talent must work in and around.

Are the constraints currently shaping your organization proportionate to the situations you now face? What range of response do they permit? What forms of initiative do they quietly discourage? What possibilities have become unthinkable—not because they were rejected, but because they never fit the field in the first place?

The architecture is already at work; what defines your leadership now is how you choose to shape it.

 Snowden, D. J., & Boone, M. E. (2007). A Leader’s Framework for Decision Making. Harvard Business Review, 85(11), 68–76.

Your middle managers sit at the intersection of every information flow in your organization. They see patterns from above and below, connect insights across departments, and understand both strategic context and operational reality.

So why do your best ideas die there?

What if the problem isn’t middle management? What if middle management is actually your secret intelligence weapon, just deployed in the wrong system?

From Ancient Agreements to Algorithmic Overlords: The Surprising Power of the Human Middle

For millennia, Homo sapiens have conquered the planet not with teeth or claws, but with an unparalleled ability to cooperate flexibly in large numbers. Our unique trick, as historian Yuval Noah Harari notes, is the creation of shared fictions—myths, religions, and, in the modern era, the limited liability corporation [1]. These fictions allow us to build intricate systems of cooperation, the most enduring of which is the hierarchy. From the legions of Rome to the factories of the Industrial Revolution, a clear chain of command was the operating system for human progress. It was simple, scalable, and brutally effective.

But the world that hierarchy was built for is gone. We have traded the predictable, mechanical rhythms of the factory floor for the chaotic, exponential pace of a networked age. And the very structure that enabled our past success is now becoming our greatest liability. The modern organization is a marvel of distributed intelligence, a vast network of human sensors processing immense volumes of information at the edges. Yet, we still try to funnel this torrent of insight through a narrow, rigid pipeline designed for a bygone era. The result? A tragic and costly suppression of the very intelligence we need to survive and thrive.

Nowhere is this paradox more acute and more misunderstood, than in the middle of our organizations. We lament that our best ideas “die in middle management.” We see this layer as a bottleneck, a frozen tundra of bureaucracy and resistance. But what if this is a profound misreading of the situation? What if the problem isn’t the people in the middle, but the system we’ve trapped them in? What if your middle managers are not a bug in the system, but a secret and powerful feature waiting to be unlocked?

The System Has a Soul, and It’s Tearing Itself Apart

Organizational theorist Barry Oshry has spent decades demonstrating a predictable and tragic pattern that emerges whenever humans organize. He calls it the dance of the “Tops, Middles, and Bottoms” [2]. Tops, burdened with ultimate accountability, feel overwhelmed by unmanageable complexity. Bottoms, on the front lines of execution, feel vulnerable and invisible, seeing problems with crystal clarity but lacking the power to enact systemic change.

And the Middles? The Middles are systematically torn apart. They are the human conduits, pulled between the strategic imperatives from above and the operational realities from below. They are tasked with translating an abstract vision into concrete reality, often with conflicting priorities and insufficient resources. This isn’t a failure of character; it is a system condition. When we fail to see this system, we blame the people caught within it. The middle manager becomes a bottleneck not by choice, but by design.

They learn, through a thousand painful cuts, to filter. They sanitize intelligence flowing upward to protect already-burdened executives from the messy, contradictory truth of the front lines. They manage expectations downward to shield already-vulnerable teams from the whiplash of strategic shifts.

This filtering, born of a desire to make the system work, slowly and inexorably kills the very intelligence the system needs. The organization creates what Stanford’s Bob Sutton and Huggy Rao call “bad friction”—layers of approval, translation, and political calculation that make sharing vital intelligence exponentially harder than keeping it local [3].

In this environment, intelligence, like water, follows the path of least resistance. And when that path leads to a dead end, your most valuable insights stay trapped in the organizational equivalent of a stagnant pond.

The Ghost in the Machine: Misreading the Signals

This systemic dysfunction leads to a dangerous form of organizational hallucination. Leaders, disconnected from the ground truth, misinterpret the signals that do manage to trickle up. Consider two stories from the front lines of organizational change:

The Overconfidence Mirage: A retail VP, preparing to roll out a massive operational change across more than a thousand stores, was briefed by his middle managers to expect resistance. The system had taught them to anticipate pushback. But when a parallel intelligence network was opened, gathering direct, unfiltered feedback from the frontline staff, a shocking and far more dangerous reality was revealed. The teams weren’t resistant; they were overconfident. They believed the changes would be simple, blind to the cascading failures their leaders could foresee. The risk wasn’t a failure to comply; it was a failure to prepare. The rollout strategy was inverted—from selling a change to co-creating a readiness plan. The result was not just a successful implementation, but the creation of genuine agency for 1,000 people who, for the first time, felt like contributors, not just cogs.

The Friction Revelation: A three-star general was convinced his civilian experts were the source of resistance to a critical national security initiative. His command structure—the middle—reported that the experts were dragging their feet. But direct, distributed sensing revealed the opposite. The experts were desperate for the change to succeed. The “resistance” was not to the what, but to the how. They were resisting the bureaucracy that forced them up and down a rigid chain of command for simple collaboration. The friction wasn’t human; it was systemic. The middle managers weren’t the problem; they were simply describing the symptoms of a disease they were powerless to cure.

In both cases, the middle was not a source of failure, but a sensor reporting on a broken system. They were faithfully transmitting the distorted signals the hierarchy was designed to produce. The tragedy is that we, as leaders, have been trained to shoot the messenger.

The Intelligence Insurgency: Rewiring the System

If hierarchy is the problem, the solution is not to tear it down. Anarchy is not a viable business model. The solution is to augment it, to build a parallel, more dynamic system for intelligence flow—a Human Sensor Network. Think of it as the organizational equivalent of the nervous system, running alongside the skeleton of the hierarchy. The hierarchy provides structure and decision rights; the network provides real-time, distributed intelligence.

And who is perfectly positioned at the critical nodes of this new network? Your middle managers. They are the biological hubs, the natural connectors who already sit at the intersection of strategic intent and operational reality. The breakthrough comes when we stop seeing them as filters and start equipping them as amplifiers.

This is not a theoretical fantasy. Research from INSEAD provides stunning validation: 80% of transformation programs led by middle managers succeed, compared to a paltry 20% of those led by senior management alone [4]. Why? Because the middle possesses the contextual knowledge, the relational trust, and the cross-functional visibility to make change real. If you think about it, they may be the only ones who can see the whole system from the inside out.

Transforming this latent potential into active capability requires a fundamental shift in our leadership operating system. We must move from a model that treats middle managers as bottlenecks to one that empowers them as hubs.

This is more than just a change in process; it is a change in philosophy. It requires that we give middle managers the strategic context they crave and the genuine authority they need. It means trusting them not just to execute a plan, but to sense and respond to a constantly changing environment. It means giving them the tools—like the distributed sensing technologies mentioned in the case studies—to bypass the friction of the formal hierarchy and tap directly into the system’s (organization’s) collective intelligence.

The Technology of Transformation: Building Your Human Sensor Network

The most advanced organizations are deploying what we call Human Sensor Networks using sophisticated sensing technologies that fundamentally alter the energy gradients of information flow. These are not traditional survey tools or dashboards that aggregate and abstract. They are designed to preserve the richness and context of distributed intelligence while making it accessible to decision-makers in real time.

Consider SenseMaker™, a technology that enables people throughout the system to share and interpret their own micro-narratives—the small stories that reveal emerging patterns invisible to conventional analytics. Unlike traditional feedback mechanisms that force people to fit their experience into predetermined categories, SenseMaker allows the organization to see itself through the eyes of those living within it. The intelligence isn’t filtered through middle management translation; it flows directly, with middle managers adding their own contextual insights rather than controlling the narrative. This was the technology we used in the military example above.

Or MassSense™, which acts like organizational sonar—rapidly gathering perspectives from across your entire system on specific situations, decisions, or changes. Instead of waiting weeks for intelligence to percolate up through the layers (by which time the world has moved on), you can actively engage what people throughout the organization are seeing and thinking right now. This was the technology we used in the retail example above.

The Choice: Industrial Age Relic or Intelligence Age Engine?

The intelligence required for your survival is likely already flowing through your system. As I wrote about in previous blogs, it is in the observations of your frontline staff, the frustrated workarounds of your engineers, and the informal networks of your admins. Your middle managers are already positioned at every critical intersection of this flow.

We can continue to operate on an Industrial Revolution blueprint, treating our middle managers as cogs in a machine, destined to become bottlenecks. We can invest millions in AI and analytics, hoping to find insights in the data while the human intelligence that provides its context withers on the vine.

Or, we can make a different choice. We can see our organizations not as employee contributors to be controlled, but as part of a complex, living system to be cultivated. We can choose to see our middles not as a problem to be solved, but as the solution we have been looking for—the critical hubs of our emerging human intelligence network.

This is not simply about being better leaders. It is about building a more sapient organization. It is about having the courage to rewire the deep structures of power and information flow. The work is not easy, but the alternative is obsolescence. The intelligence is there. The potential is there. The only question is whether you will unleash it.

Next in the series: The Constraint Paradox: Why Unlimited Freedom Kills Creativity and How to Design the Enabling Boundaries That Unleash Breakthroughs.

References

[1] Harari, Y. N. (2015). Sapiens: A Brief History of Humankind. Harper.

[2] Oshry, B. (2007). Seeing Systems: Unlocking the Mysteries of Organizational Life. Berrett-Koehler Publishers.

[3] Sutton, R. I., & Rao, H. (2024). The Friction Project: How Smart Leaders Make the Right Things Easier and the Wrong Things Harder. St. Martin’s Press.

[4] Deloitte. (2025, March 24). What’s the future of management? Deloitte Insights. https://www2.deloitte.com/us/en/insights/focus/human-capital-trends/2025/future-of-the-middle-manager.htm

l

While you’re building artificial intelligence systems to capture insights, your organizational systems are systematically filtering out the human intelligence you already have.

Consider the Challenger disaster. NASA engineers knew the O-rings would fail in cold weather. They had the data, ran the tests, and understood the risk. But their intelligence never reached the decision-makers who could have delayed the launch. The organizational system filtered out the very information that could have prevented the tragedy.

A few decades later, with the Columbia vehicle, NASA engineers knew foam strikes were damaging heat shields. But over time, the culture transformed this critical intelligence through language: “foam strikes” became “known anomalies” became “acceptable risks.” The intelligence was there, but the organizational system normalized it away until people stopped looking.

After working with thousands of leaders, I would bet my new car that your organization faces the same dynamics every day, just with lower stakes. The intelligence you need most is already in your building, but your Leadership Operating System may be unintentionally designed to suppress it rather than surface it.

The Two-Current Problem

Every organization operates with two intelligence currents running simultaneously:

The first is the surface current—the world of reports, dashboards, meetings, and official channels. This current carries what people think you want to hear, information that’s been sanitized, aggregated, and stripped of its most valuable context. It’s clean, measurable, and fits neatly into PowerPoint presentations. Most executives navigate entirely in this surface current, optimizing what gets measured and making decisions based on information that’s traveled through multiple filters.

The second is the subsurface current—the world of hallway conversations, workarounds that actually work, and insights that never make it past the first filter. This is where people share what they actually see, not what they think they’re supposed to see. It’s where solutions emerge organically, where early warning signals first appear, where the gap between “how things are supposed to work” and “how things actually work” becomes visible.

What most executives sense, but haven’t really yet accepted is the subsurface current often contains your most valuable competitive intelligence. While your competitors are looking at the same industry reports and market data you are, they don’t have access to what your EA knows about client sentiment, what your newest hire sees about industry blind spots, or what your middle manager has discovered about operational efficiency.

But there is a catch. The subsurface current is fragile and when people sense that their insights aren’t valued or that sharing what they really see might be politically dangerous, the subsurface current goes deeper underground. The intelligence doesn’t disappear entirely—it just becomes invisible to leadership. In fact, the smarter your people are, the faster they learn what not to share. They quickly figure out which insights are welcome and which ones aren’t. They learn to read the room, to sense what leadership wants to hear, and to filter their intelligence accordingly.

Here are two recurring examples I see across our client systems: someone shares a pattern they’re seeing and gets told “that’s just anecdotal” or “we need more data.” They don’t just stop sharing that particular insight—they learn to stop sharing patterns altogether. Another one is when someone creates a better way to do something and gets told to “follow the process,” they don’t just abandon that innovation—they learn to keep future innovations quiet and take their discretionary ideas to other outlets.

Over time, intelligent people become intelligent about what not to share. The very people whose insights you need most become the least likely to offer them. They develop an adaptive behavior of keeping valuable insights to themselves to avoid negative consequences - “learned intelligence suppression.”

This creates a vicious cycle. Leaders, starved of real intelligence from the subsurface current, make decisions based on incomplete information. When those decisions don’t work as expected, they often conclude they need better data or more rigorous analysis. So they add more filters, more approval processes, more formal channels—which drives the subsurface current even deeper.

We Need More Data Beta!

My partner and I enjoy skiing in the backcountry, and one thing I have learned is that in backcountry skiing, my life depends on “beta”—scattered bits of raw intelligence. A local mentions recent snowfall. Someone coming down the mountain talks about conditions on the north-facing slope or how it was “great pow all day”. Simply feeling cold walking back from the bar at night is a piece of beta about what might be beneath the surface tomorrow. Without all of these unpolished clues, I wouldn’t understand a coherent as a picture and could risk entering avalanche-prone terrain unaware.

The backcountry community has a discipline around receiving beta. In organizations, we’ve lost the art of receiving beta. Instead, we demand well-curated data, analyzed and wrapped in neat presentations, before we’ll listen. Such insistence filters out the unpolished, frontline insights that might reveal subtle risks or unexpected shifts. Our people are like knowledgeable locals whose beta gets dismissed because it doesn’t come with a PowerPoint deck.

Here are some pieces of Beta to look out for:

Early warning beta emerges from people who interact with external stakeholders—customers, suppliers, partners, competitors. They often sense shifts in sentiment, behavior, or market conditions months before those shifts show up in formal metrics. Be careful to dismiss this intelligence as “narrow”, “soft” or “subjective” just because it can’t be generalized or quantified immediately. As someone who spans industries, I often share something I learned from one into the other and watch as some discount the “beta” because simply because it is not industry specific. I never mean for it to be a conclusion, but as a single piece of beta, when combined with other single pieces of beta, may provide a more coherent picture.

Operational beta comes from those closest to the operational front lines. They know which processes really work, which systems are truly mission-critical, and where the inefficiencies hide. They’ve often developed informal solutions that work better than official procedures, but these innovations remain invisible because they exist outside formal channels. Even worse, they have to hide these innovations from you so they are not reprimanded and made to do things inefficiently. We always do “ride alongs” or “day in the life of” in the first few weeks of a client engagement to understand how things really work on the front lines.

Strategic beta exists in the minds of people who are in unique vantage points where there is unsuspecting data that does not usually register as strategically interesting. For example, the person who schedules meetings knows which initiatives are really getting attention and which ones are consistently cancelled or pushed out. The person who processes expenses sees where money is actually spent and where the bottle necks lie in approvals. The person who handles customer complaints understands which problems are systemic versus isolated. While the reportable data is available, the contextual intelligence is often invisible.

Innovation beta lives in the spaces between formal roles and departments. It emerges when people with different expertise and perspectives connect informally to solve problems or explore opportunities. It happens when people are solving problems or encountering anomalies. Because they are anomalies, by definition, there is no trend to report. These cross-functional insights often represent your biggest innovation opportunities, but they rarely surface through formal innovation processes.

As Rabindranath Tagore wrote, “The truth comes as a conqueror only because we have lost the art of receiving it as a guest.” Leaders that successfully unlock the distributed genius in their ecosystem have rediscovered this art. They realize intelligence is voluntary—people choose whether to share what they know based on whether they believe it will be valued and acted upon, no matter how small.

This doesn’t mean accepting every piece of scattered intelligence as gospel. Like avalanche beta, each fragment is just one piece of the puzzle. The skill lies in treating these hints as starting points—prompts to investigate further and validate what’s real and relevant. You maintain perspective, neither dismissing raw insights nor letting one snippet define your entire outlook. Instead of demanding certainty before listening, you learn to do what the master listener (and my mother), Jill Rickards advises, “gather everything and let it lie before you”. Or more urgently, gather intelligence as if your company’s life depends on it—because in a rapidly changing world, it does.

As you take on building the organizational capacity for emergent and distributed intelligence, understand that such intelligence operates differently than artificial intelligence. First, such intelligence is discretionary not obligatory and does not respond to demands but rather genuine invitations followed by validation and appreciation. Second, interpretation of such intelligence leverages abductive reasoning—making sense of incomplete information, forming hypotheses about what might be causing something, picking up on subtle energy or mood shifts, and inferring the best explanation from limited clues. It’s the “aha” moment when someone says “I think something’s changing because...” based on subtle signals. This is in contrast to inductive reasoning—finding patterns in historical data and making predictions based on correlations. It’s the difference between inferring what might be happening now and analyzing what already happened. Interestingly, humans excel at abductive while AI excels at inductive. When these two forms of reasoning work together, organizations can anticipate their future in ways neither could achieve alone.

The Choice

When you successfully tap into distributed genius, something remarkable happens: The intelligence compounds. People start connecting insights across departments, new patterns become visible, and cross-functional solutions become possible.

But the real transformation is cultural. When people realize their intelligence is genuinely valued—not just their compliance or execution, but their actual insights and perspectives—they shift from being passive recipients of direction to active contributors to organizational intelligence. The organization becomes more than the sum of its parts because it’s finally accessing the intelligence of all its parts. Real engagement comes from engaging! (PS – yes, this is how to get those eNPS scores to genuinely rise.)

The opportunity for each of us is to acknowledge the intelligence that is already in our buildings (virtual and physical). It’s in the observations our people make every day, the patterns they notice, the solutions they create, the insights they develop. The question isn’t whether this intelligence exists—it does.

The question is whether you’re going to invest in human intelligence at the same level you are investing in artificial intelligence. Your competitors are looking at the same market data you are. But they don’t have access to what your people know. The question is: Do you?

Next: Why your best ideas die in middle management, and how to create parallel intelligence networks that actually work.

As a Captain in the military, I had a bigger contribution to make than my company-grade rank allowed. My superiors were phenomenal; they recognized and pulled the best out of me, creating opportunities within their power. But outside my immediate chain of command, there wasn’t an outlet. The system itself couldn’t access what I had to offer.

It wasn’t anyone’s fault. Hierarchy isn’t bad—it’s actually essential for creating decision rights, clarity of direction and execution. The military’s hierarchical structure exists for good reasons and works brilliantly for what it’s designed to do. But it also meant that I could only contribute to the top of my rank. As a Captain, there was a certain scope of contribution I could make, and that was it.

I realized I was going to have to wait 10-15 years before I’d be able to contribute systemically and strategically in the ways I knew I could. So I got out.

Here’s the part that’s bothered me ever since: Soon after leaving, I started consulting for the same military organization. Suddenly, I could make a much more widespread impact from outside the system than I ever could from within it. Same person. Same insights. Same commitment. But the system could finally access what I had to offer—because I was no longer trapped by its necessary structures.

I loved my life in the military, and I think I would have really enjoyed staying in. But the system—despite being well-designed for its primary purposes—couldn’t unlock the distributed genius within its own ranks. I encounter people like this all the time: talented individuals with big contributions to make, trapped in pockets of excellence by systems that can’t access their full potential.

That experience launched nearly 20 years of work helping organizations unlock the hidden genius already in their systems. I’ve spent my career teaching executives how to recognize the intelligence that exists throughout their organizations and create conditions that enable it to flow. I know this stuff works because I’ve seen it transform companies.

So when two key people on my own team decided to move on recently, I had to face an uncomfortable truth: Even after two decades of studying this problem, I had fallen into the same trap.

The irony wasn’t lost on me. Here I was, teaching other leaders how to access distributed intelligence, while somehow failing to fully access the intelligence of my own people. The very expertise that I brought to client organizations had created a blind spot in my own leadership.

In a virtual company like mine, you don’t have the luxury of hallway conversations or chance encounters in the break room. The informal networks that carry so much of an organization’s real-time intelligence are harder to see and easier to miss. My focus on client work and the structured nature of virtual communication created a perfect storm for this kind of blind spot.

The Contradiction Nobody Talks About

Here’s the reality unfolding in organizations right now: Companies are investing record amounts in innovation programs, engagement initiatives, and artificial intelligence—while simultaneously running systems that suppress the human intelligence they already have.

Your organization probably has multiple channels for capturing intelligence: innovation workstreams, engagement surveys, town halls, skip-level meetings, brainstorming sessions, teams or slack channels, etc. You’ve invested in these systems because you genuinely want to hear from people.

But after nearly 20 years of working with organizations to unlock distributed genius, I’ve seen a consistent pattern: Critical insights take weeks to reach decision-makers—if they reach them at all. Breakthrough innovations come from the same predictable sources, not because others lack ideas, but because the system can’t access them. Your frontline manager knows exactly which process is creating bottlenecks, but raising it requires building a business case that takes more time than just working around it. Your customer service team can predict which clients are about to churn based on subtle shifts in tone, but your retention metrics only track explicit complaints. Your high-potential employee has an idea that could reshape your business model, but she’s learned that “strategic thinking” is reserved for people two levels above her.

The intelligence is there. The channels exist. But the system, through a thousand small signals about what’s valued, what’s risky, and what’s worth the political capital, is filtering it out.

This is an intelligence problem with a measurable cost.

Research from the DevOps Research and Assessment (DORA) program shows high-information-flow environments where intelligence can move freely are roughly twice as likely to meet or exceed their organizational goals in profitability, productivity and market share. The difference isn’t just cultural, it’s measurable advantage.

And profit aside, it’s also about preventing accidents, defects, and in some cases, preventable harm. Gallup’s longitudinal research across thousands of organizations shows that when people don’t feel their opinions count, safety incidents rise by 64% and quality defects increase by 41%. In healthcare, over 70% of serious patient harm events are failures of communicating the available intelligence.

The Middle Ground Most Organizations Live In

I learn something new and am awed by every single client I’ve ever worked with. When we start our work, in fact, most things are going right. Most leaders do want ideas, and most have channels for it. Most have engagement scores that really aren’t terrible. They all have smart people doing good work.

But somewhere between having the intelligence and channeling it in such a way that it can be considered and acted on, something gets lost. Ideas surface but don’t reach their full potential. People share insights but the really valuable stuff—the uncomfortable truths, the early warning signals, the innovations that challenge assumptions—those stay underground.

It’s likely that your system does not completely block intelligence. Rather, it is optimized for a certain kind of intelligence—the kind that’s understandable, repeatable, easily actionable, and fits neatly into existing frameworks and surveys. Meanwhile, the intelligence that could give you breakthrough advantage (the kind that challenges assumptions, reveals uncomfortable truths, or emerges from unexpected places), learns to stay quiet.

This intelligence is fragile, unconventional and sometimes uncomfortable. It lives in hallway conversations, informal networks, and the space between what people are supposed to say and what they actually think. When people sense that their real insights aren’t valued or that sharing what they truly see might be politically risky, that intelligence goes underground.

The cruel irony is that the smarter your people are, the faster they learn what not to share. They figure out which insights are welcome and which ones aren’t. They learn to read the room, to sense what leadership wants to hear, and to filter their intelligence accordingly.

Just like the military system couldn’t access my full contribution as a Captain, my own system wasn’t accessing the full contribution of my people. Different context, same dynamic.

If you’re like most leaders I work with, you’re dealing with some version of these challenges. Our leadership systems are intentionally designed to filter rather than amplify.

The Operating System Problem

Every organization runs on what amounts to a Leadership Operating System—the combination of strategy, culture, leadership practices, and operating structures that determines how intelligence flows and decisions get made. Most of us are running Industrial Revolution software in an exponential world.

When we realize our system isn’t working, we typically add more features to the existing infrastructure rather than upgrading the foundation. We layer engagement surveys on top of disengaging practices. We create innovation labs while maintaining approval processes that kill innovation. We launch “speak up” campaigns while keeping the hierarchical structures that taught people to stay quiet in the first place.

The intent is always good. But when the foundation is faulty, adding more sophisticated tools just creates more sophisticated dysfunction. It’s like installing a state-of-the-art navigation system in a car with a broken engine—you’ll know exactly where you’re stuck, but you still won’t be moving.

These legacy systems were designed for predictable work where intelligence could be centralized and decisions could be made at the top. They worked brilliantly when change was slow and work was routine. But they’re badly mismatched for today’s reality, where the most valuable intelligence often lives at the edges of your organization and the best solutions emerge from unexpected connections.

Most leadership systems are optimized for scale and execution, not intelligence and adaptation. This design enabled the remarkable growth and operational excellence we’ve achieved over the past century. But systems optimized for one thing inevitably create trade-offs. As an unintended consequence of that design, they filter out dissent, suppress experimentation, and reward compliance over insight. The very people whose intelligence could give you competitive advantage learn that sharing their real insights is risky or pointless— not because leaders don’t want to hear it, but because the system wasn’t built to receive it.

What’s Coming in This Series

Over the next several weeks, I’ll be exploring how to upgrade your Leadership Operating System to unlock the human intelligence your organization needs to thrive. We’ll cover:

• The Intelligence Already in Your Building - How to see and access the distributed genius that’s hiding in plain sight, and why “beta” matters more than big data

• Why Your Best Ideas Die in Middle Management - Understanding the system dynamics that trap intelligence and how to transform bottlenecks into amplification hubs

• The Constraint Paradox - Why unlimited freedom kills creativity and how to design enabling boundaries that channel intelligence toward breakthrough results

• When Surveys Conceal Lived Truth - Moving beyond reported experience to access what people actually experience, and why the gap between them contains your most valuable intelligence

• Navigating the VUCA Nexus - How to lead when the world feels volatile, uncertain, complex, and ambiguous, and how to find clarity in the chaos

• The Coherence Advantage - How to integrate everything into a Leadership Operating System that enables human and artificial intelligence to work together

A Simple Question to Ask to Open the Conversation

Moving forward, I’m committing to regularly asking my team a simple but powerful question: “What is it like to be you right now?” Not “How satisfied are you?” or “What do you need from me?” but “What is it like to be you in this organization, doing this work, at this moment?”

I suspect the answers will surprise me. They’ll show me the difference between what I think I’m creating and what people are actually experiencing. They’ll reveal intelligence about our own system that I can’t see from my vantage point.

More than anything, they’ll help me practice what I teach: that the intelligence we need most is often the intelligence we’re least likely to hear unless we create specific conditions for it to emerge. Research shows when leaders create such conditions, information is able to flow to leaders with actual decision rights. Such properly routed intelligence improves unit performance, while intelligence that only circulates among peers without agency dissipates into noise.

A Different Kind of Upgrade

You have a choice that will define your organization’s next decade: You can continue optimizing Industrial Revolution systems while investing in AI, hoping technology will compensate for suppressed human intelligence. Or you can upgrade to a Leadership Operating System designed to unlock both human and artificial intelligence working together.

It’s the same story across industries: when systems dampen voice, results suffer. When systems enable intelligence to flow (especially to those with decision rights) performance follows. The choice isn’t whether to have structure. It’s whether your structure enables or suppresses the intelligence that could give you breakthrough advantage. I’m fully diving into articulating what we have learned to address this challenge starting with this series.

Next: “The Intelligence Already in Your Building” - How to see the genius that’s been there all along

What intelligence might be hiding in plain sight in your organization? Hit reply and share your thoughts—I read every response and often explore these insights in future articles.

Subscribe now