Collective intelligence has become one of the most influential ideas in systems change.
Across governments, innovation labs, and civic platforms, a common assumption now shapes practice: bring together diverse perspectives, and systems become more legible. Patterns emerge. Decisions improve.
But there is a problem.
We are getting better at understanding systems. We are not getting better at changing them.
Many of the most advanced collective intelligence efforts, from political economy analysis to participatory governance, produce insight without producing sustained change. Systems become easier to see, yet remain difficult to shift.
This points to a deeper constraint.
The issue is not only how well systems understand themselves. It is whether actors within them can act together under conditions of uncertainty.
This is where we need to move beyond collective intelligence and towards collaborative intelligence.
Seeing systems is not the same as changing them
Collective intelligence is often understood as the ability of groups to generate insights that exceed the knowledge of any individual participant. But as Geoff Mulgan in Big Mind argues, it is better understood more broadly as the capacity of systems to combine data, knowledge, judgement, and action.
In principle, this spans the full cycle: sensing, interpreting, deciding, and responding.
In practice, collective intelligence strengthens how systems understand and decide. It makes systems more legible. It reveals patterns, surfaces constraints, and improves judgment.
This has significantly advanced systems design and governance.
But improved understanding does not, on its own, change how systems behave.
Systemic change requires actors with different roles, incentives, and authority to engage in ways that reshape interactions over time. This is where many efforts stall.
Political economy analysis, for example, can map power, incentives, and informal networks with increasing precision. Yet these insights rarely translate into durable reform.
The constraint is not analytical. It is behavioural.
Systems struggle to convert shared understanding into coordinated action under conditions of uncertainty, disagreement, and shifting incentives.
Collective intelligence improves how systems think. It does not determine how actors engage, experiment, or adapt in practice.
From collective intelligence to collaborative intelligence
This is where a different capability becomes necessary.
Dawna Markova and Angie McArthur describe collaborative intelligence as the capacity of diverse individuals to integrate their different cognitive strengths into collective thinking. This provides an important foundation, highlighting how diversity can deepen shared understanding.
But in complex systems, and particularly in political environments, the central challenge is not only cognitive. It is relational and behavioural.
Collaborative intelligence can therefore be understood more precisely as: The capacity of diverse actors to act within a system together through iterative experimentation, and to learn from how the system responds.
This is where the work of Reos Partners becomes particularly instructive. Their stretch collaboration approach shows that actors do not need agreement, prior trust, or shared control to work together. They can begin with action.
Collaboration, in this sense, is not alignment. It is joint experimentation.
Actors take small, deliberate steps, not to implement predefined solutions, but to test how the system behaves. These actions function as probes, generating signals:
· which actors engage, and under what conditions
· where resistance strengthens or weakens
· which informal rules hold, and which begin to shift
· what new possibilities emerge through interaction
Through this process, collaboration becomes a way of learning from the system itself.
Actors do not need full agreement in advance. They learn how to move together by acting and adapting as the system responds.
An important consequence of this process is how trust develops.
In complex systems, trust is rarely a starting point. It emerges through shared experience by observing how others act under pressure, how commitments are honoured, and how risks are taken.
Over time, repeated interaction builds a different kind of foundation: trust grounded in behaviour, not alignment.
Collaborative intelligence is therefore not an extension of collective intelligence. It is a distinct capability: the ability to navigate uncertainty, work across difference, and generate new patterns of behaviour through interaction.
Collective vs Collaborative Intelligence
The distinction becomes clearer when viewed through a simple framework.
Why systems change requires collaborative intelligence
This distinction has become visible in practice. In my governance work in the Democratic Republic of Congo, political economy analysis was redesigned as a collaborative sense-making process rather than a diagnostic exercise.
Actors from ministries, civil society, and frontline services mapped their experiences using a shared framework. Patterns that had appeared isolated became recognisable as part of a wider system.
This was collective intelligence: the system became able to see itself.
But the shift that mattered came later.
When participants formed small cross-system groups to experiment together, the work changed. The question was no longer whether the system was understood, but whether actors could act within it in ways that generated new information.
These groups became the early architecture of collaborative intelligence capable of observing, probing, and adapting in response to how the system behaved.
A missing distinction in the systems change field
The systems change community has invested heavily in collective intelligence. This is necessary. No single actor can understand complex systems alone.
But understanding is not the endpoint.
Once systems become more legible, the central challenge shifts. It is no longer about insight. It is about how actors engage with uncertainty, difference, and power through action.
This is not primarily an intelligence problem.
It is a problem of interaction.
The next frontier in systems change lies in strengthening collaborative intelligence: the capacity to experiment, sense, and adapt within complex systems.
From collaborative intelligence to collaborative political intelligence
In political systems, collaboration does not occur in neutral conditions. Power shapes what is possible. Authority shapes who can act. Incentives shape who participates.
For this reason, for new, durable political systems to emerge, collaborative intelligence must evolve into collaborative political intelligence.
This is the ability of actors not only to experiment together, but to do so in ways that navigate power, negotiate interests, and reshape how authority and coordination function in practice.
At this level, It is no longer a question of whether actors can act together. It is whether their interactions begin to shift the incentive structures that govern behaviour.
This is what determines whether change endures.
Many collaborative efforts generate insight, experimentation, and short-term gains. But these often remain fragile. When attention shifts, systems revert.
Durable change requires something more specific.
It emerges when interactions produce self-reinforcing patterns of behaviour across actors.
In these conditions, expectations shift, norms stabilise, and deviation becomes costly. Not because it is centrally enforced, but because it disrupts relationships and practices that actors now depend on.
This is what durability looks like in practice: not sustained reform, not institutionalisation, but a self-reinforcing behavioural equilibrium.
From this perspective, collaborative political intelligence is not about better coordination. It is about acting in ways that make new patterns of behaviour more stable than the old ones.
This is where the work becomes most challenging and most consequential.
It is also the question explored in my next article.