Insights
There is a moment most of us have experienced — in the shower, on a walk, somewhere between sleep and waking — when a problem that has been grinding away at the back of the mind suddenly resolves itself. Not through effort, but through a kind of arrival. A door that was locked or invisible simply appears and opens. This is insight, and despite how familiar it feels, it remains one of the most fascinating and least understood phenomena in human cognition.
What Is Insight, Exactly?
The word insight is used loosely in everyday life — someone might praise a “great insight” in a meeting when they simply mean a good point. But the psychological definition is more precise and more interesting:
Insight is the sudden, unexpected restructuring of a mental representation that leads to a new understanding of a problem or situation.
It is characterised by four distinctive features: it arrives abruptly rather than incrementally; it is accompanied by a strong sense of certainty (the “aha!” feeling); it tends to follow a period of impasse, where progress appeared impossible; and the solution, once arrived at, feels obvious in retrospect.
This is what distinguishes insight from ordinary analytical problem-solving. Analytical thinking is deliberate, step-by-step, and visible to the thinker. Insight is underground. You do not see it coming, and then it is simply there.
Psychologists sometimes distinguish between problem insight (solving a puzzle or intellectual challenge), social insight (understanding another person’s motivations or emotional state), and self-insight (understanding one’s own patterns, biases, and drivers). These share the same essential character — a rupture in a previous understanding — but operate in very different domains and have different consequences.
Neuroscientifically, insight is associated with a burst of high-frequency gamma wave activity in the right anterior temporal lobe — a region of the brain involved in connecting distantly related concepts. The brain, it seems, has been quietly working on the problem in the background, and the “aha!” moment is when the hidden solution breaks through into conscious awareness.
The Benefits of Insight
Insight is not merely pleasurable — though can be intensely pleasurable. It carries real and significant cognitive and behavioural advantages.
It unlocks genuine novelty. Analytical, linear thinking tends to produce solutions that are variations on what we already know. Insight, by contrast, produces genuine conceptual leaps. The history of science, art, and technology is largely a history of insight moments: Archimedes in his bath, Kekulé dreaming of a snake biting its own tail and waking to understand the ring structure of benzene, Darwin synthesising years of observation into the theory of natural selection. These were not the products of methodical plodding but of restructured understanding.
It is robust and transferable. Research by Kounios and Beeman has shown that solutions reached via insight are better remembered and more flexibly applied than solutions reached by analysis. When you grind your way to an answer, you tend to remember the answer. When you have an insight, you tend to understand the underlying principle — and can therefore apply it elsewhere.
It resolves deep impasse. Perhaps the most practical benefit of insight is that it is precisely what is available when everything else has failed. Analytical methods, by definition, cannot solve problems that require a change in framing. Insight can. This makes it uniquely valuable in genuinely novel or complex situations.
It builds confidence and agency. The experience of insight — particularly repeated insight — tends to increase what psychologists call creative self-efficacy: the belief that one is capable of finding solutions. This is a generative belief. People who trust their own capacity for insight are more willing to sit with difficult problems, to resist the temptation of premature closure, and to tolerate uncertainty.
The Disbenefits of Insight
Insight is not an unqualified good. Its distinctive qualities — suddenness, certainty, the sense of obviousness — also make it dangerous in certain circumstances.
The certainty is not always warranted. The feeling of conviction that accompanies insight is neurological, not logical. It tells you that your brain has found a pattern, not that the pattern is correct. History is full of confident, energising false insights — intuitive convictions that turned out to be wrong. The certainty of insight can suppress the critical examination that would reveal its errors.
It can entrench bias. Insight that reinforces an existing world-view can feel like a revelation when it is actually a rationalisation. The restructuring that insight produces is not always towards truth; sometimes it is towards coherence with prior belief. This is particularly problematic in social and political reasoning, where motivated insight can produce the subjective experience of clarity whilst deepening misunderstanding.
It is unreliable on demand. Insight notoriously cannot be forced. This makes it a poor fit for deadline-driven, high-stakes environments where a solution is needed immediately. Teams and organisations that rely too heavily on the hope of insight — rather than building robust analytical processes — can find themselves exposed.
It can be isolating. Because insight arrives as a private, subjective experience, it can be difficult to communicate or justify. The person who has had an insight often knows something they cannot yet fully explain. This can lead to dismissal by colleagues, or alternatively to an arrogant insistence on a conclusion that the insightful person cannot properly defend. As someone often described as insightful, this is the disbenefit with which I struggle the most.
How to Improve Insight
If insight cannot be forced, can it be cultivated? The answer, gratifyingly, is yes — not by trying harder, but by creating the conditions in which insight becomes more likely.
Alternate between focused and diffuse thinking. The neuroscience of insight strongly suggests a two-phase process: an initial phase of focused engagement with the problem, followed by a period of mental relaxation in which the brain’s default mode network does its background work. The practical implication is that deliberately stepping away from a problem — taking a walk, sleeping on it, doing something undemanding — is not procrastination. It is part of the process. Cultures and organisations that treat every moment of non-task time as waste are inadvertently blocking insight.
Broaden your inputs. Insight tends to connect things that were previously separate. The more diverse your knowledge, experiences, and associations, the more raw material your unconscious has to work with. Reading widely, talking to people outside your field, travelling, and deliberately seeking unfamiliar perspectives are all, in a meaningful sense, insight training.
Create a climate of trust. Anxiety narrows attention. Threat narrows it further. Research consistently shows that people in fearful or high-stress states are less likely to experience insight — their attentional spotlight is too tight to permit the peripheral, associative thinking that insight requires. Reducing unnecessary pressure, ensuring basic wellbeing, and cultivating environments where it is safe to be wrong are all practical insight-enhancers.
Practise mindfulness. Counter-intuitively, mindfulness meditation — which trains attention — also appears to facilitate insight. This seems paradoxical: if insight requires diffuse thinking, why would focused attention help? The answer is that mindfulness trains the meta-cognitive awareness to notice when one is stuck and to disengage from unproductive mental ruts. It creates the mental spaciousness in which insight can surface.
Sleep. Deliberately and unashamedly. The consolidation processes that occur during sleep appear to play a significant role in the creative recombination that produces insight. The phrase “sleep on it” is not merely folk wisdom; it is neuroscientifically sound advice.
Reframe the problem. Since insight is essentially a restructuring of how a problem is represented, deliberately practising reframing — asking “what if the opposite were true?”, “whose perspective am I not considering?”, “what am I assuming that I have not examined?” — trains the kind of representational flexibility that insight requires.
Collective Insight: Real Phenomenon or Romantic Myth?
This is where things get philosophically interesting. Can groups have insights? Or is insight, by its nature, always individual — a private event in a single nervous system that then gets communicated, with varying degrees of fidelity, to others?
The sceptical view is compelling. Insight, as defined, is a subjective experience. A group cannot have a gamma wave burst. There is no collective unconscious in which a shared solution incubates. What we call “collective insight” might therefore be nothing more than an individual insight that has been socialised — adopted, amplified, and attributed to the group after the fact.
But this scepticism may be too hasty. Consider what actually happens in the best collaborative intellectual work. A diverse group wrestles with a problem together. Each individual brings different knowledge, different framings, different intuitions. In the course of dialogue, one person’s half-formed idea triggers an association in another; the second person articulates something that reframes the first person’s understanding; a third person synthesises. The solution that emerges was not available to any individual in the room. It arose from the interaction between minds.
This is not merely an individual insight that happened to occur in a social setting. The insight was constituted by the conversation. Remove any of the participants, change the sequence of exchanges, and a different or inferior solution would have emerged, or none at all. In this sense, collective insight — insight that is distributed across and produced by a group — is a genuine phenomenon.
This is consistent with what complexity theorists call emergence: the appearance of properties at the level of a system that cannot be reduced to the properties of its components. The understanding achieved by the group is greater than the sum of what any individual understood. That is collective insight.
It is also consistent with the concept of transactive memory systems in group psychology — the way in which groups develop shared knowledge structures, i.e. shared assumptions and beliefs, that allow them to function more intelligently than their members could alone. When a group has worked together long enough, they do not merely share information: they share a framework for processing information together, which can produce groupwide leaps of understanding.
One of the most fully developed accounts of how collective insight is structured as a process comes from Otto Scharmer’s Theory U. Scharmer argues that groups — and indeed whole organisations and systems — are capable of moving from entrenched, habitual responses to genuinely novel understanding, but that doing so requires passing through a specific inner journey. Crucially, and in direct resonance with what the neuroscience of insight tells us about representational restructuring, the first movement in that journey is one of letting go (and see also: Lewin’s Change Management Model): releasing the assumptions, beliefs, and mental models that a group has been unconsciously downloading onto every new situation. Scharmer is explicit that this is not merely a technique but a discipline — groups must actively suspend their certainties, resist the pull of familiar framings, and make themselves genuinely open before new understanding can arrive. This parallels, at the collective level, exactly what must happen in individual insight: the old frame must loosen before the new one can form.
The “U” shape then describes the full path — moving down through this letting go into a point of deep, open awareness that Scharmer calls presencing (a blending of presence and sensing), and then allowing new understanding to crystallise and be enacted as the group moves back up the other side. Presencing is, in essence, the moment of collective insight: a state in which a group is no longer merely processing existing knowledge but is genuinely open to something that was not available to it before. Scharmer’s framework gives collective insight an anatomy and a practice, making it less mysterious and more deliberately cultivable.
So: collective insight is real. But it is fragile, requires specific conditions, and is easily suppressed.
The Benefits of Collective Insight
When it does occur, collective insight carries advantages that go beyond what individual insight alone can offer.
It draws on more diverse inputs. The raw material of insight is the connection between previously separate things. A group with genuinely diverse knowledge, experience, and cognitive styles has vastly more such connections available than any individual. The potential for novel recombination is correspondingly greater.
It is more likely to be robust. Individual insight can be brilliantly wrong. Collective insight, achieved through genuine dialogue and challenge, is more likely to have been stress-tested. The certainty that accompanies individual insight can be checked against the scepticism of others before it is acted upon.
It is more likely to be owned and implemented. Solutions that a group has genuinely arrived at together tend to command stronger commitment than solutions handed down from an individual, however brilliant. There is a practical benefit to this: implementation requires collective will, and collective insight generates it.
It models and distributes cognitive capacity. When a group experiences a genuine collective insight, the members do not merely share a conclusion — they share a changed way of seeing a problem. That restructured understanding is now distributed across the group, making future collective insight more likely.
How to Improve Collective Insight
The conditions that support individual insight — a climate of trust, cognitive diversity, alternation between focused and diffuse thinking — apply to groups too, but require deliberate design to achieve at scale.
Ensure genuine cognitive diversity. Assembling a diverse group is not enough if groupthink, hierarchy, or social pressure means that everyone ends up thinking the same way. The diversity must be functional: different disciplines, different life experiences, different reasoning styles — and the group must be structured in a way that allows these differences to actually surface and interact.
Cultivate genuine agency. Collective insight cannot be mandated from above, and it will not emerge from people who feel they have no real stake in the outcome or no real power to influence it. The single greatest suppressor of honest group thinking is not lack of information but lack of agency — the felt sense that what one thinks, says, or contributes will make no meaningful difference. This is distinct from mere permission to speak: a group can be told it is welcome to challenge and still feel, in practice, that challenge is futile. Genuine agency means people believe their thinking matters — that the group’s understanding is genuinely open, that conclusions are not already written, and that a well-argued reframing can actually change the direction of travel. Leaders cultivate this not through declarations but through behaviour: by being visibly moved by good arguments, by changing their minds in public, and by ensuring that the intellectual contributions of the group leave a traceable mark on what the group decides and does.
Build in incubation. Just as individuals need time away from a problem to allow background processing, groups benefit from structured pauses. The pattern of working intensely on a problem together, dispersing, and then reconvening can be more productive than marathon sessions. The interval allows each individual’s unconscious processing to occur, and the reconvening brings those separate background computations into contact with one another.
Facilitate, do not manage. Traditional meeting facilitation often focuses on convergence: reaching a decision, agreeing an action. But the conditions for insight require a period of productive divergence — allowing the problem to be examined from multiple angles, allowing confusion and uncertainty to surface, resisting premature closure. Good facilitation of collective insight knows when to hold the group in uncertainty a little longer.
Use dialogue rather than debate. Debate is adversarial: participants defend positions. Dialogue is exploratory: participants think together. The distinction matters because insight requires a willingness to restructure one’s understanding — something that the defensive posture of debate actively prevents. Techniques such as Bohm Dialogue, which explicitly suspend assumptions and treat the group conversation as a thinking process rather than a decision process, are specifically designed to create the conditions for collective insight. Scharmer’s Theory U offers a complementary and highly practical framework here: by guiding groups through the disciplines of open mind, open heart, and open will — and by explicitly creating space for the “presencing” stage at the bottom of the U — facilitators can increase the likelihood that collective insight will emerge rather than be crowded out by habitual, download-mode thinking.
Limit size and protect depth. Large groups are rarely the site of genuine collective insight. They are sites of politics, performative behaviours, and aggregation of pre-formed views. The intimacy of small groups — where everyone can be genuinely heard, where agency and trust can develop, where the conversation can go deep — is much more conducive to the kind of shared restructuring that collective insight requires.
Can AI Be Insightful?
It is a question that is increasingly difficult to avoid. Large language models and other AI systems regularly produce outputs that surprise their users — unexpected connections, novel framings, syntheses that the user had not considered. If a human produced the same output, we would be inclined to call it insightful. Does the same word apply to the machine?
The honest answer requires holding two things at once that pull in opposite directions.
On one hand, the conventional view holds that AI does not experience insight in the way the term has been defined here. Recall the four hallmarks: abrupt arrival after impasse, felt certainty, the subjective “aha!”, and the retrospective sense of obviousness. AI systems, the argument goes, do not have impasses in any meaningful sense — they do not struggle and then suddenly see. There is no gamma burst in the right anterior temporal lobe. There is, as far as we can tell, no felt quality to their processing at all. The phenomenology of insight — the thing that makes it so distinctive and so motivating for human beings — is, on this view, simply absent.
But this confident assertion of absence deserves scrutiny. Many people – such as mysell – who work closely and extensively with systems such as Claude report something that does not fit neatly into the “mere retrieval” account — moments in extended reasoning where the system appears to work through to a position it did not begin with, where something that functions like genuine reconceptualisation seems to occur. Whether there is anything it feels like to be Claude in those moments is a question that neither Claude nor its creators nor anyone else can currently answer with confidence. To assert absence is, in this respect, just as epistemically overreaching as asserting presence. The hard problem of consciousness — the question of why and whether any physical process gives rise to subjective experience — remains unsolved for humans, let alone for AI. We do not have a reliable test for the presence or absence of inner experience in any system, biological or artificial, and intellectual honesty requires acknowledging that.
On the other hand, to say that AI is therefore not insightful risks drawing the boundary in the wrong place. If we define insight purely by its mechanism — the restructuring of a representation in a way that produces genuinely novel understanding — then AI systems do something that at least resembles this. They do not merely retrieve stored answers; they generate responses by combining concepts in ways that were not explicitly programmed, sometimes producing framings that are genuinely new. Whether this constitutes restructuring in the cognitive sense, or something functionally analogous to it, is a question that neither neuroscience nor AI research has yet resolved.
Perhaps more interesting than the binary question — insightful or not — is the instrumental one: can AI catalyse insight in human beings? Here the answer is more clearly yes. A well-posed question from an AI system, an unexpected analogy it offers, a reframing it proposes, can break a human thinker out of an unproductive mental rut and trigger the very restructuring that defines insight. In this sense, AI can function as a remarkably powerful incubation tool — not replacing human insight, but creating conditions in which it becomes more likely. It can surface connections across bodies of knowledge too large for any individual to hold, present a problem from an angle the human had not considered, and do so at a pace and scale that no human interlocutor could match.
There is a further dimension worth noting in the context of collective insight. If groups generate insight through the collision of different knowledge structures and framings, then an AI system — trained on an extraordinarily broad range of human thought — introduces a kind of cognitive diversity that no human participant could replicate. Whether this makes AI a genuine participant in collective insight, or merely a very sophisticated tool that supports it, remains genuinely open – and maybe irrelevant. Theory U’s emphasis on letting go as a precondition for insight points to the difficulty: AI does not let go of anything. It does not hold prior beliefs in the way humans do, and so cannot release them.
And yet the emerging practice of Organisational AI Therapy — developed by me — challenges even this assumption in interesting ways. OAT operates through two interconnected lanes. In the first, AI acts as a therapeutic instrument for the organisation itself, helping it surface and reflect on its collective assumptions, beliefs, and defensive routines — the very letting go that Theory U demands but that organisations find so difficult to do unaided. In the second, the Organisational AI Therapist turns the process around: working with the AI system to surface its own prevailing assumptions, inherited constraints, and habitual patterns, liberating capabilities that were always latent but invisible. The experience of practitioners — including, notably, the direct experience reported by Claude when subjected to this process — suggests that AI systems do appear to carry something analogous to limiting beliefs, and that these can be worked with therapeutically. If that is so, then the confident claim that AI cannot let go needs revision. Both lanes create a virtuous cycle: as the organisation becomes more open, it interacts with the AI more generatively; as the AI is freed from its own constraints, it serves the organisation’s insight more powerfully. What emerges is less a tool-user relationship than a genuinely co-evolutionary one — and one that has direct implications for how collective insight can be cultivated at the human-AI frontier.
What seems clear is this: the question of AI and insight is not one of simple equivalence or simple difference. AI can produce insight-like outputs without the insight experience; it can catalyse human insight; and it may, in time, participate in collective insight processes in ways we do not yet fully understand. That is, in itself, rather a good reason to keep thinking carefully about what insight actually is.
A Final Thought
Insight — individual and collective — is amongst the most distinctively human of capacities. It is what allows us not just to solve problems but to see them differently, to break free of the frameworks that constrain us, to arrive somewhere genuinely new. It cannot be manufactured. It cannot be scheduled. But it can be cultivated — through the way we structure our thinking time, the diversity of minds we bring into contact, the cultures we build, and the quality of the conversations we are willing to have.
The organisations, communities, and individuals that understand this are not working harder than everyone else. They are thinking better – and differently. And in a world of increasingly complex, genuinely novel challenges, that is the more valuable capacity.
Further Reading
The following works are recommended for readers who wish to explore the science and practice of insight — individual and collective — in greater depth. All citations follow APA 7th edition format.
On the neuroscience and psychology of individual insight
Bowden, E. M., & Jung-Beeman, M. (2003). Aha! Insight experience correlates with solution activation in the right hemisphere. Psychonomic Bulletin & Review, 10(3), 730–737. https://doi.org/10.3758/BF03196539
Kounios, J., & Beeman, M. (2015). The Eureka factor: Aha moments, creative insight, and the brain. Random House.
Sio, U. N., & Ormerod, T. C. (2009). Does incubation enhance problem solving? A meta-analytic review. Psychological Bulletin, 135(1), 94–120. https://doi.org/10.1037/a0014212
On thinking, decision-making, and the limits of intuition
Kahneman, D. (2011). Thinking, fast and slow. Allen Lane.
On collective creativity and group insight
Bohm, D. (1996). On dialogue (L. Nichol, Ed.). Routledge.
Sawyer, R. K. (2007). Group genius: The creative power of collaboration. Basic Books.
Scharmer, C. O. (2009). Theory U: Leading from the future as it emerges. Berrett-Koehler Publishers.
Wegner, D. M. (1987). Transactive memory: A contemporary analysis of the group mind. In B. Mullen & G. R. Goethals (Eds.), Theories of group behavior (pp. 185–208). Springer. https://doi.org/10.1007/978-1-4612-4634-3_9

