Over the past few years, working on governance reform, systems change, collaboration, and public sector innovation, I have kept encountering the same tension.
We generate insight.
We design new approaches.
We convene the right people.
And yet, this rarely translates into sustained change in how systems operate. At a time of overlapping crises, addressing this gap is becoming harder to ignore.
This piece is an attempt to explore a pattern I’m starting to notice, drawing from my experience in fragile political economies and from learning alongside those working on systems-led collaboration, governance architecture, and innovation in public systems.
Seeing the system: Political economy as civic infrastructure
In my recent series on governance reform in the DRC, I explored how political economy analysis (PEA) might be reimagined, not as a report, but as a form of civic infrastructure.
By turning PEA into a shared sensemaking process, ministries, inspectors, and civil society actors who rarely shared a space began to view the system together. Through storytelling and collaborative analysis, challenges that had been experienced as local or individual were recognised as connected and systemic. Patterns converged around leverage points that no single actor could see alone.
What emerged wasn’t just a sharper diagnosis. It was a shared understanding of how power, incentives, and informal rules were shaping outcomes. A foundation for thinking and working politically (TWP).
However, whilst Political economy helped those within the system see themselves more clearly, it did not, on its own, establish new ways of deciding, coordinating, or acting together. It revealed how the system worked, without yet reshaping the structures or relationships through which collective action could occur.
In other words, political economy made the system legible, but it did not help redesign the pathways through which people could act on that understanding.
What collaboration and design can’t do on their own
Once political economy made the system legible, it is natural to turn toward collaboration and governance architecture or design as ways to act on that insight.
Collaboration creates spaces where people from across institutions and positions can come together, test interpretations, and begin to build trust. Governance architecture offers ways of imagining new decision pathways, coordination mechanisms, and institutional arrangements that might work better than the ones already in place.
Both approaches are essential but limited on their own.
Collaboration is often asked to carry too much weight. It can generate conditions for trust and learning, but without changes to how authority, resources, and decisions are structured, that momentum struggles to translate into durable shifts. Over time, collaboration risks becoming a holding space for frustration rather than a driver of change.
Governance design faces a different challenge. New architectures and institutional reforms can be thoughtfully conceived, but without ongoing engagement with how power actually operates, design can drift into abstraction. Decision pathways look plausible on paper, but remain disconnected from the incentives and relationships that shape behaviour in practice.
What I began to notice is that collaboration and design tend to struggle at the same point: when political dynamics shift, resistance emerges, or informal power overrides formal arrangements.
At that point, insight needs to be refreshed, strategies need to adapt, and ways of working need to change, not once, but repeatedly. This is the work of thinking and working politically: staying attentive to power, incentives, and shifting relationships while action is already underway.
Systemic change doesn’t fail because any one of these approaches is wrong. It falters because none of them is sufficient on its own, and because we rarely design them to work together, over time.
This is where the question stopped being “How do we collaborate better?” or “How do we design better institutions?” and became something more fundamental:
How do people keep learning and adapting politically while trying to change how a system works?
What begins to matter in practice
As I’ve been working through these questions, reflecting on early experimentation, conversations, and parallel work across systems change, a pattern has begun to emerge.
When systems move beyond diagnosis and begin to shift, what seems to matter less is any single intervention, and more how a small number of underlying capacities interact. Capacities that allow people to see the system more clearly, act together within it, and adapt as power dynamics and conditions change.
At this stage, I’m exploring whether three such capacities consistently show up when systems begin to shift, and why they seem to matter most when they are held together as mutually reinforcing ways of working, rather than applied in isolation.
- Governance architecture – When insight has no pathway to influence decisions
Governance architecture refers to how decisions, authority, and coordination are structured. One thing that becomes increasingly visible once action is attempted is whether there are real pathways through which insight can influence decisions.
Governance architecture refers to the institutional and relational pathways through which decisions are made, authority is exercised, and collective action becomes possible. This includes formal structures, but also the informal rules, routines, and permissions that shape how the system actually functions.
In fragile political economies, institutions often exist on paper, while real decision-making and coordination take place through parallel or informal pathways. When governance is treated primarily as an organisational form or technical design, it can struggle to engage with this lived reality.
Without governance architecture, insight and collaboration remain disconnected from how decisions are taken, resources are allocated, and authority is exercised. Understanding exists, relationships form, but there is no clear route for either to shape the system’s core operating logic.
From this perspective, governance architecture is less about prescribing solutions and more about shaping the conditions under which new forms of coordination, learning, and decision-making might plausibly emerge.
- Systems-led collaboration – When collaboration creates energy but not leverage
Systems-led collaboration refers to how diverse actors work across differences over time. In complex and fragile systems, authority is distributed, incentives are misaligned, and no single actor is able to move the system alone. Systems-led collaboration points to the capacity for people with different interests, levels of power, and institutional positions to engage with one another over time.
This is not about alignment or consensus. It is about creating enough relational continuity for people to test ideas, negotiate trade-offs, and learn from how the system responds.
Collaboration, in this sense, is not simply a delivery mechanism for pre-designed solutions. It is one of the primary ways systems learn in practice. Through action, disagreement, resistance, and feedback, collaboration can surface political dynamics that are difficult to see in advance.
At the same time, without changes to how authority and decision-making are structured, collaboration risks generating energy without leverage. Conversations deepen, understanding grows, but systems tend to revert once attention shifts or pressure mounts.
This suggests that collaboration may be necessary for change, but not sufficient on its own.
- Thinking and working politically – Staying oriented as power and incentives shift
As collaboration and design efforts unfold, political dynamics rarely remain static. Thinking and working politically (TWP) refers to the capacity to stay attentive to power, incentives, and relationships as action is already underway, and to adapt accordingly. Rather than treating political economy as a one-off diagnostic, it frames political insight as something that must be continually updated through practice.
In many reform and innovation efforts, political analysis is conducted upfront, while learning and adaptation are left implicit or informal. In these conditions, collaboration and governance design can end up being built on partial or outdated pictures of reality.
Without the capacity to think and work politically, governance architecture risks being designed for a system that no longer exists, and collaboration risks reinforcing familiar power dynamics rather than shifting them.
At the same time, this capacity does not sit apart from the others. It is generated through collaboration and becomes consequential only when there are governance pathways through which insight can influence decisions and reshape coordination.
From this perspective, thinking and working politically is less an input to change than a continuous capability, a way of staying grounded in the system as it changes.
A dynamic learning system, not a sequence
What matters most here is not the labels but the logic of mutual dependency:
· Governance architecture without collaboration or political intelligence risks being designed for a system that exists on paper, not in practice.
· Collaboration without governance architecture risks generating trust, energy, and insight without any pathway to influence decisions or shift outcomes.
· Thinking and working politically without collaboration or architecture risks remaining descriptive, sharp in analysis, but weak in its ability to reshape how the system actually works.
When any one of these capacities is missing, systems tend to stall, fragment, or revert to familiar patterns. When they reinforce one another, systemic change, even if slow, contested, and uneven, becomes more plausible.
This is not a conclusion but an invitation to sense-making
I want to be clear: this is not a claim about what works, nor a framework others should adopt.
It is a working hypothesis, an attempt to make sense of what appears to matter when systems transition from understanding to action, and from action to adaptation.
Based on my observations, I have not seen a single organisation, discipline, or methodology holding all three of these capacities equally well. Instead, it seems that these capacities are distributed across various actors and practices. However, if they were to begin to interact, looser constellations of practice can form networks that learn, coordinate, and gradually build legitimacy through action rather than formal mandate.
Whether these capacities are sufficient, how they interact in practice, and what it would take to intentionally weave them together remains an open question, and one I’m continuing to explore through practice and dialogue.
If you’ve read this far, I hope you’re also curious about what becomes possible when governance architecture, systems-led collaboration, and thinking and working politically are held together over time.
Over the coming months, I’m convening a small sensemaking conversation with practitioners working across these domains to test this proposition, challenge it, and refine it together.
If this resonates with your work, I would welcome the conversation.