The Mosaic and the Phalanx: What the US-Iran Conflict Reveals About Organizational Design and Resilience

Living in Greenwich Village in September 2001, I watched the Twin Towers fall from my Carmine Street rooftop. For much of the rest of that academic year, I tried to make sense of what had happened on that Tuesday morning, for myself and my NYU students. During the same period, both the organizational management and security studies communities were already theorizing the changes to their core subjects under the pressure of the unprecedented event.

In the spring of 2003, I attended an early “Un-Conference” in New York and encountered a non-traditional format that was itself a performance of the decentralized principles being discussed. I was also not entirely surprised when a presenter proposed, with genuine intellectual seriousness, that Al Qaeda’s organizational design pointed toward the future of network design. I recall the room went quiet in the way rooms only go quiet when something is simultaneously obvious and unspeakable (at least not yet, not there).

The argument was not a moral endorsement; rather, it was a structural observation, drawing on the recent, pre-9/11 work of RAND analysts (2001), that argued how a leaderless, cellular, “all-channel” network could absorb devastating strikes to its center and continue to function (Arquilla and Ronfeldt, 2001). Although several people walked out, others stayed and argued well beyond the session’s scheduled ending with no one fully agreeing, but no one entirely dismissing the argument either. That tension between the efficiency of hierarchy and the resilience of modularity has persisted over my succeeding two decades of work with leaders and organizations. And now, twenty-three years on, it is being foregrounded, in blood and fire, on two entirely different continents.

I. Beyond Resilience

Traditional organizations are arguably built like a phalanx, with disciplined ranks and interlocked functions, made to be formidable when advancing on familiar terrain. As American classicist and military historian Victor Davis Hanson writes, the classical Greek and Macedonian phalanx was one of history’s most effective systems of coordinated force, being nearly unbreakable from the front, capable of shattering almost anything in its path on level ground (Hanson, 2000). From an organizational perspective, its structural logic arguably passed into the first generation of modern management articulated by Frederick Winslow Taylor. The early twentieth-century engineer and pioneering management thinker’s Principles of Scientific Management standardized every movement of every worker on the same principle of disciplined collective execution, substituting central design for local variation and concentrating power through tight coordination (Taylor, 1911).

The flaw in both systems, as the ancient Greek historian Polybius identified when analyzing Rome’s defeat of the Macedonian phalanx at Cynoscephalae in 197 BC, was structural. He understood that the phalanx suffered from flanking exposure and terrain dependence and risked cascading collapse wherever the formation broke (Polybius, 1926). What proved true of the Spartan line has proven equally true of many bureaucratic firms, that the same tight integration that concentrated frontal power made the whole formation vulnerable to an adversary willing to engage from an unexpected angle.

The standard approach to overcoming this flaw has been advanced by the former U.S. Army general in charge of fighting the Taliban and Al Qaeda in Afghanistan, Stanley McChrystal (and his colleagues), in Team of Teams (McChrystal et al., 2015). His response was to build the capacity of the military system to absorb shocks and return to prior form – resilience– through distributed communication and shared awareness and knowledge. That prescription, sound as far as it goes, has since become something close to a leadership cliché, appearing in everything from annual reports to social media posts, often with diminishing analytical content.

Polymath and writer Nassim Taleb has explored a harder concept that is also worth noting here. In Antifragile: Things That Gain From Disorder, Taleb drew a sharp line between systems that merely survive disorder and systems that actually gain from it and emerge from the shock stronger and more capable than before (Taleb, 2012). Resilience, in his account, represents a lower bar, which means you bounce back to where you started. Anti-fragility means the disruption becomes an input to improvement, the way muscles grow through stress or immune systems sharpen through exposure.

The distinction has a clear correlate in business leadership and organizations. Netflix’s Chaos Engineering program deliberately injects system failures, through an aptly named “Chaos Monkey” tool, into production infrastructure to force engineers to build services that improve rather than merely survive disruption (Basiri et al., 2016). The company does not hope its systems will withstand failures; instead, the program engineers internalize processes and operations to improve because of those failures. For organizational designers, the distinction matters enormously, because what looks like resilience is sometimes just concealed rigidity, and what looks like vulnerability can, under the right structural conditions, turn out to be something altogether more generative.

II. Two Doctrines, One Logic

Considering resilience in light of the current conflict in the Gulf, this is where the story becomes (at least for me) both unexpected and instructive. Two deeply opposed military and political traditions have, through entirely separate reasoning, arrived at the same doctrinal and design conclusion.

The first is the Iranian “Mosaic Defensive Strategy,” a doctrine codified nearly two decades ago and stress-tested publicly since late February 2026, when the U.S. Operation Epic Fury eliminated close to forty senior Iranian leaders, including the Supreme Leader and high-ranking IRGC commanders (Bezhan, 2026). On any conventional organizational logic, this killing of senior leadership should have produced collapse by decapitating a hierarchical system at its apex.

Instead, Iran’s 31 provincial Iranian Revolutionary Guard Corps commands, structured as autonomous units with their own resources, intelligence, and pre-delegated authority, continued to operate on what Al Jazeera described as “general instructions given in advance” rather than real-time orders from a center that no longer existed as it had (Al Jazeera, 2026). The Iranian state did not collapse because it had, by design, embedded its intent into the structure itself rather than concentrating its intent in any single node. That is precisely the cellular logic that Connell identified in his analysis of Iranian military doctrine decades ago and before the doctrine was tested under fire (Connell 2010).

The Iranian model is resilient in the technical sense that it survived a decapitating strike, though the evidence thus far suggests it falls short of being anti-fragile in Taleb’s sense. Absent central oversight, autonomous provincial units have been responsible for the targeting of neutral vessels in the Gulf of Oman, uncoordinated actions that a functioning center might have prevented (The Soufan Center, 2026). Though the system has survived and continued, it remains unclear, as of this writing, if, besides local enhancements to targeting, it has improved overall. In an exceptional series of brief analyses, AI strategist Matthew Kilbane argues the U.S.-Israeli-Iranian conflict is, in fact, the “anti-fragile war” (Kilbane, 2026).

A second tradition is represented by the U.S. military’s DARPA’s Mosaic Warfare, which may move closer toward anti-fragility. This model was conceived not in the shadow of geopolitical vulnerability but of economic asymmetry, when a single aircraft can cost hundreds of millions of dollars and its destruction constitutes a strategic crisis. DARPA’s solution was to decompose the traditional military platform into thousands of low-cost, AI-linked tiles, separating sensing, shooting, and communicating into distributed functions that collectively overwhelm an adversary’s decision-making (Deptula, et al., 2019).

In principle, the U.S. system is designed not merely to survive strikes but to proliferate faster than an adversary can respond, gaining effectiveness as it scales (though the extent to which it has been implemented, and actually converted pressure into offensive capability, is an open question). By contrast, the Iranian model appears to be operational and to have absorbed actual strikes successfully (at least). Both represent mosaic thinking, though the extent to which they are resilient or may even approximate genuine anti-fragility remains to be seen.

III. Convergent Evolution and What It Signals

That ideologically opposed systems converged independently on the same structural logic is a signal of the leadership, strategic, and technological realities of our time. Convergent evolution in biology indicates environmental pressure powerful enough to produce identical solutions through entirely different lineages. The organizational equivalent is an environment so complex and volatile that hierarchy, for all its coordination advantages in stable conditions, cannot absorb shocks quickly enough to survive.

We can see that platform-era businesses arrived at the same conclusion through market pressure rather than military necessity, even though the organizational instinct for flexibility long preceded the digital era. Business historian Alfred D. Chandler, Jr. analyzed the first great corporate unbundling of the functional phalanx in his groundbreaking analysis, Strategy and Structure. Studying the extraordinary growth of firms like DuPont, General Motors, and Sears, Chandler saw how executives dismantled centralized functional organization and replaced it with semi-autonomous divisions, each carrying its own operating logic, coordinated by a lean corporate center (Chandler, 1962).

Alfred Sloan’s multidivisional GM of the 1940s and 1950s is not yet a mosaic, but, as discussed by Chandler, it is the historical midpoint when large organizations first broke the rigid hierarchical formation in response to environmental complexity. Decades later, around 2002, Jeff Bezos pushed the logic further. At Amazon, the API mandate required every team to expose its capabilities as a service to every other team, creating a modular internal architecture in which no single failure could cascade across the whole (Yegge, 2011).

The U.S. military recently reached the same structural conclusion. A 2019 Department of the Army publication puts this concisely.

In practice, mission command tends to be decentralized, informal, and flexible. Plans, orders, and graphics should be as simple and concise as possible, designed for maximum flexibility during execution. By decentralizing decision-making authority, mission command increases tempo and improves subordinates’ abilities to act quickly in fluid and chaotic situations (U.S. Department of the Army, 2019, p. 1-22).

It is telling to read such a formal acknowledgment, from within an institution that once built its operational identity around phalanx-like coordination, that centralized control and decision-making has been superseded. Whether in military doctrine, industrial organization, or cloud computing, it is the same pressure producing these designs – environments that punish single points of failure and reward systems capable of reconfiguring under stress.

IV. The Leader as Curator

The mosaic model demands from its leaders a role that is deceptively simple to describe and operationally demanding to execute. In business, Zhang Ruimin’s transformation of Haier over two decades offers the clearest business illustration. During that time, the CEO dismantled one of China’s largest appliance companies and rebuilt it as approximately 4,000 micro-enterprises, each carrying its own P&L, hiring authority, and autonomous strategic decision-making, within a shared enabling platform. He called the model, “RenDanHeYi,” which literally linked employee value directly to user value (Fischer, Lago & Liu, 2013).

In a McKinsey Quarterly interview, Zhang explained concisely the leadership required in this structure: “You must delegate all those powers – decision-making, hiring and firing, and setting compensation – to the microenterprises themselves. Giving up control is actually an important part of the model” (McKinsey Quarterly, 2021). His account of his own legacy completes the curator image: “changing a hierarchical and bureaucratic enterprise into an ecosystem, changing a whole garden into a rain forest.”

Conversely, the traditional chess-player model, in which the leader moves pieces according to a centrally held plan, assumes both that the board is visible and that the pieces will stay where placed. Neither assumption holds in today’s genuinely turbulent environments. Strategic management pioneer Henry Mintzberg formulated a structural typology that precisely names the organizational logic operating here. Among his five “configurations,” he contrasted the “machine bureaucracy,” which is centralized, standardized, optimized for stable and predictable conditions, and the “adhocracy,” which coordinates through mutual adjustment among distributed units and is designed for environments too complex for standardized rules to govern (Mintzberg, 1984).

The phalanx and the machine bureaucracy belong to the same structural family, as do the mosaic and the adhocracy. The task of the leader as curator, as the case of Zhang Ruimin and Haier makes concrete, is to ensure that the tiles of the mosaic are interoperable, oriented toward the same image, and held by “grout” strong enough to maintain the pattern when individual tiles are removed or replaced. Crucially, the grout in Haier’s model is not culture in the soft sense but platform infrastructure in the hard sense, comprising shared data systems, financial protocols, and a philosophy sufficiently internalized that the center can step back without the picture fragmenting.

This is neither passive leadership nor retreat from responsibility. Pre-delegating authority requires extraordinary clarity about organizational and strategic intent that needs to be clearer, in many ways, than command-and-control demands because there is no center available to subsequently correct for ambiguity. Zhang Ruimin’s model took more than a decade of iterative rollout to reach operational maturity (Fischer, Lago & Liu, 2013). The Iranian military doctrine likewise required nearly two decades of doctrinal preparation before it was activated. Neither happened quickly or easily, and no account of mosaic leadership is honest that omits that long commitment and preparation cost.

V. How Much Center Do You Give Up?

The mosaic model poses an operative question for every leader. It is not whether to decentralize, since twenty-first century VUCA or BANI environmental conditions appear to have settled debates about the benefits of flexibility and modularity in most settings. The question, rather, is how far to go in decentralizing before coherence collapses. The Iranian, Haier, and Netflix cases together suggest an answer organized around three design imperatives that bear directly on practice.

Intent must be pre-embedded in norms and not simply in rules (which give way under novel conditions) and in judgment formed through sustained shared experience and practice. Instead of being handed a manual in February 2026, IRGC provincial commanders acted on the basis of having spent years internalizing a doctrine. Zhang Ruimin’s micro-enterprise leaders at Haier did not simply receive P&L freedom; they operated within a philosophy absorbed through iterative engagement over more than a decade. The connective tissue of shared mental models, protocols, and purpose is the primary organizational infrastructure, not the chart, and must be treated as such in resource allocation and leadership attention.

The mosaic design must include feedback loops from the periphery back to a reconstituting center, however light. This is where the Haier model provides the clearest structural answer to what the Iranian model apparently cannot offer. Haier’s platform aggregates what the micro-enterprises learn from their direct user contact and routes that intelligence back into the ecosystem, reinforcing overall learning and improvement. Though the center does not manage operations, it learns from them and, through the transfer of knowledge, evolves the conditions for all the micro-enterprises.

A pure mosaic, as the conflict in the Gulf currently appears to demonstrate, can preserve the organization while degrading its coherence. The anti-fragile aspiration, in which distributed units not only survive but improve through turbulence and disorder, requires a center both light enough to survive its own removal and substantive enough, when it reconstitutes, to incorporate what the tiles learned while operating alone.

Finally, the leaders curating mosaic organizations must shift their metrics from tile performance to grout strength. Most management systems measure the output of individual units (the equivalent of appraising individual mosaic tiles for color or texture) rather than the integrity of the connections between them. Organizations that build the explicit measurement of cross-unit coordination, knowledge transfer, and shared situational awareness into their performance systems are designing for the volatile and uncertain environment that currently exists. Those that do not are still organizing more like phalanxes, which, while formidable from the front, are exposed on every other side and left to hope their flanks hold against today’s unpredictable threats.

References

Al Jazeera (2026, March 10) “The ‘Fourth Successor’: How Iran Planned to Fight a Long War with the US and Israel”; https://www.aljazeera.com/features/2026/3/10/the-fourth-successor-how-iran-planned-to-fight-a-long-war-with-the-us-and-israel

John Arquilla and David Ronfeldt, eds. (2001) Networks and Netwars: The Future of Terror, Crime, and Militancy,RAND; https://www.rand.org/pubs/monograph_reports/MR1382.html

Ali Basiri, Niosha Behnam, Ruud de Rooij, Lorin Hochstein, Luke Kosewski, and Justin Reynolds (2016) “Chaos Engineering,” IEEE Software, 33(3), 35–41; https://doi.org/10.1109/MS.2016.60

Frud Bezhan (2026, March 7) “With Top Brass Dead, Iran Deploys Decemtralized ‘Mosaic’ Strategy to Boost Defenses,” Radio FreeEurope/Radio Liberty; https://www.rferl.org/a/iran-irgc-israel-us-war/33697690.html

Alfred D. Chandler, Jr. (1962) Strategy and Structure: Chapters in the History of the American Industrial Enterprise, MIT Press.

Michael Connell (2010, October 11) “Iran’s Military Doctrine,” The Iran Primer, United States Institute of Peace; https://iranprimer.usip.org/resource/irans-military-doctrine

David A. Deptula and Heather R. Penney, with Lawrence Stutzriem and Mark A. Grunziger (2019) Restoring America’s Military Competitiveness: Mosaic Warfare, Mitchell Institute for Aerospace Studies, Air Force Association; https://www.mitchellaerospacepower.org/restoring-americas-military-competitiveness-mosaic-warfare/

Bill Fischer, Umberto Lago, and Fang Liu (2013). Reinventing Giants: How Chinese Global Competitor Haier Has Changed the Way Big Companies Transform, Jossey-Bass.

Viktor Davis Hanson (2000). The Western Way of War: Infantry Battle in Classical Greece, 2nd ed., University of California Press.

Matthew Kilbane (2026, March 20) “1 of 5 – The Antifragile War: The Iranian Mosaic Doctrine, Agentic AI, and How to Prevent a Perpetual War,” LinkedIn; https://www.linkedin.com/pulse/antifragile-war-iranian-mosaic-doctrine-agentic-ai-how-kilbane-0tjce/

Stanley McChrystal, Tantum Collins, David Silverman, and Chris Fussell (2015). Team of Teams: New Rules of Engagement for a Complex World, Portfolio/Penguin.

Henry Mintzberg (1984) “A Typology of Organizational Structure,” Organizations: A Quantum View, eds. Danny Miller and Peter H. Friesen, Prentice-Hall.

Polybius (1926) Histories, Fragments of Book XVIII, trans. W. R. Paton, Loeb Classical Library, Harvard University Press; https://penelope.uchicago.edu/Thayer/E/Roman/Texts/Polybius/18*.html

Nassim Nicholas Taleb (2012) Antifragile: Things that Gain from Disorder, Random House.

Frederick Winslow Taylor (1911). The Principles of Scientific Management, Harper & Row, Project Gutenberg; https://gutenberg.org/ebooks/6435

The Soufan Center (2026, March 9) “Iran’s ‘Mosaic Defense’ Strategy: Decentralization as Resilience Factor”; https://thesoufancenter.org/intelbrief-2026-march-9a/

U.S. Department of the Army (2019) ADP 6-0: Mission Command: Command and Control of Army Forces, Army Publishing Directorate;

https://armypubs.army.mil/epubs/DR_pubs/DR_a/ARN34403-ADP_6-0-000-WEB-3.pdf

Steve Yegge (2011, October 11) “Stevey’s Google Platforms Rant” [Leaked internal memo, preserved on GitHub]; Link

Zhang Ruimin. (2021, July 27) “Shattering the Status Quo: A Conversation with Haier’s Zhang Ruimin,” McKinsey Quarterly; https://www.mckinsey.com/capabilities/people-and-organizational-performance/our-insights/shattering-the-status-quo-a-conversation-with-haiers-zhang-ruimin