Introduction

Systems Thinking often promises a holistic lens through which to view complex global challenges. The transition, however, from theoretical "iceberg models" to tangible institutional change is fraught with political and structural friction. This article synthesizes a deep-dive conversation between Ranjit Menon, a transdisciplinary design researcher and systems thinking educator, and Noah Fang, an expert in enterprise architecture. Noah and Ranjit held this discussion in Apolitical’s Systems Thinking Community, exchanging comments over five weeks. Their discussion offers a sobering yet vital roadmap for those working within government. The primary takeaway is that technology is never just technical; it is a manifestation of power and software philosophy. Together, they explore the messy reality of applying systemic design.

Theme 1: Sources of Innovation – Frameworks vs. Evidence

Ranjit Menon: The Need for Depth

Systems Thinking often stalls at simplified, consensus-building frameworks like the somewhat popular "Iceberg Model". While these metaphors are useful in diverse stakeholder settings as an evolutionary, emergent process in co-design, they risk being an end in itself – they often replicate the same levels/hierarchies that are part of the “wickedness”, while also obscuring more leverage oriented approaches like system dynamics. In the public sector, dilemmas are "deeply wicked," perhaps requiring an additional tool to paper-based explorations. Integrating AI enhanced computational social science into classic System Dynamic models could compartmentalize complexity into manageable "chunks". Some tools like the Cynefin framework as a "sparring partner" can reveal important hidden interdependencies such as power relations. It’s important to note the "bulldozing" effect of leapfrogging digital public infrastructure in places like India, where technocracy and ad-hoc design can move behavioral economics in both positive and not so positive ways. One approach could be a certain "applied rigor" in systems modeling that accounts for the social and behavioral dimensions into an adaptive feedback loop.

Noah Fang: The Primacy of Politics and Research

Institutional change is rarely driven by scientific rigor or frameworks; it is primarily driven by politics. We should not start with a framework if the goal is genuine understanding. Instead, one must start with deep research to develop a "framing" specific to the context. As digital technologies and AI automate processes, organizational politics are shifting from "emotional politics" to "computational politics". When analyzing failures, we should use a three-layered analysis: the technical landscape, the organizational landscape (decisions and information flow), and the sociocultural governance landscape. Understanding today's organizations requires understanding the "software philosophies" (like object-oriented paradigms) that now dictate how these entities function.

Theme 2: Institutionalization of Innovation

Ranjit Menon: The Power of Boundary Objects

Core issues in information sharing are often buried deep within power structures rather than digital interfaces. Systems maps and what Dr. Birger Sevaldson calls "GigaMaps" are essential "boundary objects"—communication media/tools that facilitate engagement among diverse stakeholders. However, the "corruption of co-design" can also take place, when the rush for “consensus-building” overpowers the "art of disagreement". While stakeholders may "seemingly" agree on concepts during participatory workshops, the hierarchy's preferences often reassert themselves at the end of the process. “Semi-autonomous" external agencies could bypass these bureaucratic stalemates through more radical creativity and independence, acting as a dynamic force that internal cultures can no longer sustain.

Noah Fang: The Sandbox Trap

External structures are not always a panacea. Without effective "self-correcting mechanisms" or oversight, no force—internal or external—can create a lasting impact. Take the "sandbox trap": innovation hubs, agile teams, and design thinking projects often look perfect within a controlled environment, but they fail to scale because the sandbox is created at the discretion of the existing power structure. While digital structures can scale easily, social processes do not. The "scale up" myth is actually a status-quo-keeping enterprise, grounded in the naive belief that the system will automatically adapt to a good idea. When institutions cannot afford change effectively, the result is usually activism or collapse.

Theme 3: Government Orgs as Information Brokers

Noah Fang: The Law of Boundary Enablement

Organisational forms are progressing: "The orgs of the past were physical. The orgs of the present are digital. The orgs of the future will be social". This is explained through the "law of boundary enablement," which suggests that an enterprise's protection, coordination, and communication are enabled by its physical, digital, and social boundaries. Today’s organizations function like software; their processes and cultures are implemented through digital environments. Governments struggle to grasp how modern organizations work because decision-makers lack expertise in software architecture and history. While past workers were "laborers" (physical) and current workers are "knowledge workers" (digital), the future is in "relationship workers" (social).

Ranjit Menon: The Invisible Cogs

Invisible Cogs are the hidden markers behind the reality of the public sector in regions like India, where organizations remain overwhelmingly physical and excruciatingly bureaucratic with hypersocial powerplay by circumstance, not choice. There is a recent paradox too: while service delivery to citizens is being "leapfrogged" through digital infrastructure, the internal workings of government remain "pre-digital" and low-resource. This leads to the use of tools like WhatsApp for official internal and external communications, effectively making an "outdated" physical bureaucracy "social" by necessity and highly vulnerable. “Ghost cogs" in the machine include also the gatekeepers, middlemen, and consultancies—all constantly operating behind the scenes. According to Dr. Wolfgang Jonas, professor for design theory, these are in the form of rules, judgments, standards of value, etc., that are implicitly embedded in the machine’s performance. These characters and technocracy maps are often invisible on organization charts but heavily mediate the technical and organizational landscape, making it incredibly difficult to bring them into a legitimate environment for co-design.

Takeaways

• Avoid the "Sandbox Trap": More robust systems modelling and external innovation hubs are still rather limited if they are not integrated into the broader power structure with robust self-correcting mechanisms. Participatory design is often hollowed out by existing power structures that only allow change on their terms.
• Acknowledge the Politics: Institutional change is driven by political will and distinct physical, technical and social boundaries, not just better models. Practitioners must understand the ongoing systemic shift from "emotional" toward the "computational" politics and toward "relationship work."
• Focus on Social Scaling: While software scales with the click of a button, social and behavioral change requires deep, context-specific work. To truly innovate, public servants must move beyond "co-designing for consensus" and start designing for the complex, often contradictory, realities of power.
• Mapping the "Invisible Cogs": The digital infrastructures designed for service delivery should be seen as major leverages toward both desirable and undesirable change at Scale. However the invisible cogs behind such infrastructures, in the form of gatekeepers, consultants, and embedded value standards ("Ghost Cogs") still mediate the system, and these technocracy maps are to be accounted for as well - it is not enough to launch a seemingly well designed digital service alone.

Systemic design is not just about achieving a harmonious "consensus," but about managing the complex, often contradictory, realities of power. For an existing system to truly evolve with "systems intelligence", its change agents must have cross-disciplinary attitudes of being part-architect, part-social scientist, and part-technologist, prepared to confront "professional narrow-mindedness" behind the public sector machinery.

Join the conversation

To participate in conversations like this, join the Systems Thinking community.

Join the community

Gemini was used to aid in synthesizing the community conversation and drafting parts of this article.