Perspectives
Thinking on collective sensemaking and deliberative intelligence.
Articles and arguments from Hunome. Organised by what you are trying to understand — not by when something was published.
The Category
What collective sensemaking and deliberative intelligence are. Start here if the category is new to you.
What is deliberative intelligence — and why it is not what IBM means by it
The term is being claimed by data automation. Here is what it actually means — and why the distinction matters for every organisation trying to make decisions that hold.
What is collective sensemaking — and why it is different from collaboration
Collaboration produces outputs. Collective sensemaking produces understanding. The difference is not a matter of degree. It is a matter of what you are trying to build.
Collective intelligence collects. Collective sensemaking builds. Why the distinction matters.
Innovation platforms collect ideas. Prediction markets aggregate judgments. Collective intelligence is real and valuable. It is not the same as collective sensemaking.
Epistemic diversity — the organisational capability most tools are designed to eliminate
Epistemic diversity is the most undervalued and most systematically suppressed resource in organisational decision-making.
Assisted serendipity — the conditions for unexpected connection
Serendipity sounds like accident. On a well-designed deliberative platform, it is structural — the predictable outcome of bringing epistemically diverse perspectives into genuine contact with each other.
What a platform has to do to earn the word deliberation
The word deliberation is being applied to consultation platforms, discussion forums and AI-facilitated brainstorming. In most cases the word is doing work the platform is not. Five questions that separate deliberation from discussion.
The Unheard Room: assisted serendipity and the intelligence of varied minds
There is a kind of insight nobody in the room was looking for — a perspective from outside the domain that reframes what everyone had treated as a technical question. It is serendipity with structure, and it requires deliberate architecture rather than luck.
Can AI deliberate? Why the question matters more than the answer
Researchers are now asking whether large language models can replicate deliberation. The answer is instructive. But the more important question is what deliberation is actually for — and why that cannot be automated.
Most organisations are working on yesterday's knowledge, and do not know it
The knowledge lifecycle runs from anecdotal to legacy. Most AI and knowledge management tools operate near the end. The real competitive terrain is at the beginning, where understanding is still forming and the challenge and the perspectives to make sense of it are not yet clear.
AI mediation found common ground, and showed exactly where deliberation begins
A language model put between people who disagreed produced statements they endorsed more than a human mediator's. The result is real. What it measured is acceptability. Acceptability is a property of an output, not of what a group understands.
'Deliberative AI' is not deliberation. Why the difference decides whether your organisation gets smarter
The word deliberation is being borrowed to describe AI that plans and reasons. That is a real engineering idea, and a misleading name. Deliberation is something people do together, and letting a machine meaning take the word costs you the one capability that compounds.
Your enterprise software now has a feature masquerading as a deliberation feature. It is not.
Every major enterprise platform has shipped something called deliberation. None of it is. Here is how to tell the difference, and what it costs to choose the proxy.
Human-aware as a category
Human-centric keeps humans in view. Human-aware navigates what is real about them, including the dimension that observation alone cannot reach. Here is why that difference defines a category.
What human-aware means, and why it is not the same as human-centred
Human-centred design puts people at the start of the process. Human-aware thinking keeps people inside the reasoning throughout. The difference determines what you end up with.
Where will new knowledge come from when the models have read everything?
Models trained increasingly on model-shaped output converge, and the research now calls the endpoint knowledge collapse. What thins out is not the average but the tail, and the tail is where anything new was always sitting.
World model simulation for collective sensemaking and intelligence™: how Hunome builds the intelligence that compounds
Most deliberation is an event. Hunome's is a living system: one that builds continuously, accumulates with each cycle, and eventually produces something no workshop has ever produced: a collectively-held model of the world that reads what the world knows and tells you where you stand.
The Evidence
What the category produces that nothing else can. Named cases and demonstrated outcomes.
What 36 people from across the world understood together about demographic change — that no research brief could have specified
A global deliberation. Seven emergent clusters. None pre-specified. This is what collective sensemaking produces that no other method can.
What Nokia taught us about the thinking that exists before the categories do
The most consequential strategic insights do not come from analytical synthesis. They come from practices that operate in the space before the categories are set.
How a think tank found the connections between EU energy policy and nature loss that disciplinary isolation had missed
When contributors with expertise in energy economics, environmental policy, biodiversity governance and social sustainability worked in one deliberative space with a shared SparkMap on EU economic and energy policy, they surfaced a connection that sequential consultation had not reached: the mechanisms driving energy transition in EU policy were, under certain conditions, accelerating nature loss rather than reducing the pressures on it.
What Boeing and Challenger teach us about thinking in silos — and why it is now an organisational emergency
When specialists cannot see across disciplinary boundaries, complex systems develop invisible fault lines. Two of the most consequential failures of the 20th century have the same explanation.
The innovation that nobody wanted — what Segway and Google Glass teach us about sensemaking
The most expensive innovation failures share a structural cause. The organisations that produced them were not missing data. They were missing context. And context cannot be automated.
A thousand citizens' assemblies, and no operating system
The evidence that deliberative processes work is strong. What almost none of them leave behind is the understanding that produced the recommendations, so every assembly starts at zero, a thousand times over.
How would you know the deliberation worked?
Satisfaction scores measure how the day felt, which is why they stay high while nothing changes. The signals that matter are structural, and none of them can be collected on the day.
The Argument
Why existing tools fail and what changes when you replace them. For readers who feel the problem.
What AI cannot access — and why it matters for your most important decisions
AI is the most sophisticated tool ever built for processing what humans have already said. That is its strength. It is also its structural limit.
Surveys, workshops, and AI — why the tools we use to think together are designed to destroy understanding
The problem is not that these tools are poorly designed. The problem is that they are designed for something other than what organisations actually need.
The problem with knowledge that expires when the meeting ends
Every organisation has invested in collective thinking. Most of that investment evaporates within weeks. Here is why — and what living understanding looks like instead.
Deliberative intelligence as democratic infrastructure — why Europe needs collective sensemaking at scale
The failure of collective sensemaking is not only an organisational problem. It is a democratic one. And what it takes to address it exists.
Why strategy fails at execution — and what shared understanding actually means
The initiatives that stall do so for a predictable reason. The people expected to execute them had no access to the reasoning behind them.
Hunome and your AI investment — the layer that makes the difference
AI is the most sophisticated tool ever built for processing existing knowledge. Collective sensemaking is how you produce the knowledge that AI has not yet seen.
Why the platforms designed to connect us are making collective understanding harder
Social platforms were built to maximise engagement. Engagement and understanding are not the same thing — and optimising for one destroys the conditions for the other.
Culture doesn't block collective sensemaking. It determines where it starts.
The common assumption is that hierarchical or execution-driven cultures simply cannot do collective sensemaking. That is the wrong frame. Culture does not determine whether an organisation can build understanding together — it determines where that work starts.
We don't have time for this — and what that sentence is really saying
Take the objection seriously first: the people saying it are genuinely stretched. But the time is being spent either way. The real question is whether it is building anything that lasts beyond the meeting it was spent in.
AI operates from what humans have already expressed — and where that leaves your most important questions
AI is the most powerful tool ever built for working within the space of what has been documented. That is its strength. It is also the precise shape of its limit.
The problem with how organisations think together is getting worse — not better with AI
A convergence of recent research reveals a structural crisis in collective intelligence that no current tool adequately addresses.
The deadly pull to the status quo: why certainty tools produce mediocrity
Under real uncertainty organisations reach for the tools that offer the most reassurance: benchmarks, comparable cases, documented best practice. Those tools are rigorous. They also systematically select against the thinking that produces differentiation.
Generative AI has made your people faster. It has not made your organisation smarter.
What your AI investment will never solve, and what deliberative intelligence actually is.
When machines do the thinking, humans must do more of the judging
As AI automates the front end of decision-making, the competitive edge shifts to the back end — the quality of the questions, the rigour of interpretation, and the judgment that decides what to do with everything the machine produces.
The invisible groupthink — how AI is creating the conformity no safeguard can see
Organisations have decades of practice countering groupthink — devil's advocates, red teams, pre-mortems. Every one of them detects social signals. When convergence arrives through everyone consulting the same model, there are no social signals to detect.
The deliberative divide — why some organisations will outthink the rest
An emerging gap is opening between organisations that invest in deliberative capacity and those that do not. The gap compounds. The organisations on the wrong side of it rarely know which side they are on.
Emergence is not retrieval — what deliberation produces that AI cannot
AI and deliberation are deployed against the same surface problem and produce outputs that look alike. They are not the same act. One recombines what exists; the other creates what did not. The difference determines whether the decisions built on them hold.
Your AI investment is working. Your decisions aren't getting better. Here is why.
Boards are allocating at scale to AI agents, RAG pipelines, and proprietary data strategies. Every one of those investments rests on an assumption buried in the pitch deck. The assumption is wrong. And the gap it hides is where most of the value — and the risk — actually lives.
Get your deliberative house in order — and the stranger the better
Before the platform, before the SparkMap, before any of it — there is a muscle. Most organisations have let it atrophy. Here is how to find it again, and why the oddest people in your orbit are your most important asset.
People are, frankly, a lot — and this is what that costs you
There is a fantasy at the centre of most digital transformation programmes. It is rarely stated directly. But if you have ever sat in enough boardrooms, you have heard it between the lines. It goes roughly: if only we could run this without so many people in the way.
Humanity built the most powerful knowledge tools in history. The old dimension persists unnecessarily.
Three decades of extraordinary knowledge tools, all of them built in the extraction dimension. The dimension where knowledge comes into existence through deliberation has no operating system at scale, and it is where the most consequential human challenges actually live. AI acceleration is widening the gap, not closing it.
The workshop happened. The report exists. Nothing changed. Here is the structural reason why.
Workshops produce alignment on the day, and it is perishable. Reports contain conclusions without the understanding that produced them. Neither was designed to build a shared picture that persists, which is why the same themes keep recurring, cycle after cycle.
How to actually know what your organisation thinks, not what it says in meetings
There is a version of every meeting leadership sees, and a version that happens afterwards. Most organisations run almost entirely on the first, mistaking visible output for genuine understanding. What it takes to produce a picture that is characterised, relational and still developing, rather than a summary of what was said.
Involving 500 people in a decision without it becoming a tick-box exercise
Town halls inform. Surveys flatten. Focus groups do not scale. Genuine large-scale involvement needs three things at once: scale without flattening, structure without prescription, speed without selection, and no standard consultation format delivers them together.
Before the merger: the understanding gap that integration plans do not address
Due diligence covers financials, contracts, intellectual property and increasingly culture. It does not cover what the two organisations actually understand about the business they are entering together.
Cognitive surrender: what happens to your judgment when the answer arrives first
The research on AI and critical thinking splits by how you use it, not how much. Passive use tracks with judgment decay; structured use tracks with the opposite. The difficulty is that both feel identical from the inside.
How to know if your organisation understands your strategy, or is just saying it does
People say they understand the strategy and the engagement surveys agree. Expressed alignment is the rational answer when saying no costs more than saying yes.
Getting experts to build on each other instead of presenting in sequence
The conference model puts experts in sequence and calls it collaboration. What each expert holds is not what becomes available when those perspectives are genuinely in contact with each other.
The difference between buy-in and compliance, and why most change programmes only achieve one
Compliance is people doing what they are asked without resisting. Commitment is people bringing their judgment to it. Most change programmes are built to produce the first and hope for the second.
Genuine diversity of thought, without the process that takes six months and flattens it
The choice appears to be a narrow picture quickly or a broad one too late. It is a false choice, and the reason is structural: collecting diverse perspectives is not the same as preserving what makes them distinct.
Innovation processes that produce ideas you have not already thought of
Ideation is excellent at rapidly exploring the ideas a group already has the mental models to generate. Producing ideas outside that space is a different problem needing a different architecture.
ROU before ROI: you need a winning Return on Understanding for any Return on Investment
Every investment case in your organisation is built on a number nobody has priced. Before the transformation programme, before the AI stack, before the strategy refresh, there is a quieter question that decides whether any of it pays back. Did the people who must act build a shared, well-formed picture of what they are acting on? That is Return on Understanding. It gates every Return on Investment you will approve this year.
Smart people. Good data. Wrong decision. The structural explanation.
The organisations that make consequential strategic mistakes are not short of intelligence, data or rigour. The failure is architectural: the knowledge that would have changed the decision existed, and the process was not built to receive it.
Strategy that does not expire: how to plan when the world is changing faster than your cycle
The planning cycle assumed the key assumptions would hold for twelve to thirty-six months. For most organisations they no longer do, and adding scenarios does not fix it.
The people who already knew: why organisations keep discovering risk after the fact
After every significant failure the same conversation happens: the people closest to the problem could see it coming. This is not a communication failure. It is a structural one, and it recurs with remarkable consistency across organisations and industries.
What a genuine public consultation actually produces, and what has to change to get there
Public consultations are run with genuine intent and still are not taken seriously, by the people asked to take part or by the institutions that commission them. The cause is structural, not presentational.
Why leadership teams think alike, and what it costs the rest of the organisation
The most capable leadership teams are often the ones most at risk from intellectual homogeneity. Hiring selected for it, socialisation reinforced it, and meeting culture rewards it. It feels like strategic clarity. Frequently it is strategic risk, invisible from the inside.