After every significant organisational failure there is a version of the same conversation: the product launch that missed the market, the integration that fell apart, the strategic initiative that stalled within twelve months. It emerges in post-mortems and exit interviews and the honest conversations that happen once the pressure is off. It goes roughly like this: the people closest to the problem could see it coming. They tried, in various ways, to surface it. The signal did not reach the decision level. The decision was made without it.
This is not a communication failure in any simple sense. It is a structural one. And it recurs with remarkable consistency across organisations, industries, and types of decision.
What the signal looks like before the failure
The people who already know a risk are rarely holding a fully articulated argument. They are holding something less tidy: a feeling that something is off, a piece of specific knowledge that contradicts an assumption in the plan, a pattern they have observed that does not fit the model.
This kind of knowing is fragile in organisational settings. It is not the kind of thing that travels well through the official channels of escalation and reporting. It gets lost in translation. It gets filtered by managers who do not want to bring bad news upward. It gets suppressed by the person who holds it, who has learned that challenging the official narrative costs more than staying quiet. It arrives, if it arrives at all, too late and too attenuated to land.
The closest analogy is a geological survey. The instruments that detect early movement are sensitive, distributed, and often reading a signal that official channels are not designed to receive. Most organisations have no geological instruments. They wait for the earthquake.
The risk was visible. It was held by people who could not surface it through the channels available to them.
Why the loudest voices are not the most important ones
The people most likely to surface concerns through official channels are those with the social capital to absorb the risk of doing so: typically senior, confident, already-heard voices. These are rarely the people closest to the operational reality of the problem.
The person who knows why the product is going to miss the market is often the one who talks to customers every day. The person who knows the integration is going to break down is the one who understands the cultural difference between the two organisations at the level where work actually happens. The person who knows the strategic assumption is wrong is the specialist whose domain knowledge was not in the room when the decision was made.
None of these people are structurally positioned to get their knowledge into the decision process at the moment it would matter. The organisational hearing apparatus is not calibrated for them.
What organisations need instead
The problem is not that organisations need better escalation processes or more psychologically safe cultures, though both help. The structural problem is that the process used to gather the input that decisions are based on is designed to hear from the people who are already being heard, and is not designed to surface the distributed, fragile, early signal that lives in the organisation and does not travel through official channels.
What is needed is a process that: - Does not require social capital to participate in: the risk of contribution is low - Operates asynchronously, so people contribute when they have genuine thought rather than in scheduled sessions under observation - Receives contributions in their particular form, without forcing them into a survey question or a meeting agenda item - Maps contributions in relation to each other, so the pattern that no individual holds becomes visible
This is what deliberation at scale produces. Not the collection of opinions, but the surfacing of distributed knowledge in a form that makes its structure visible to decision-makers.
What Hunome makes possible
On a Hunome SparkMap, a strategic question is opened to the people who hold knowledge relevant to it, including those who would never raise their hand in a meeting. They contribute asynchronously, in their own terms, without the social calculation of the observed room. The AI analyses contributions as they arrive: identifying where a perspective from operations challenges an assumption from strategy, where a cluster of concern is forming below the level that formal reporting would surface, where the pattern across many small observations adds up to something significant.
The result is that leadership sees the signal before it becomes a problem. Not because people are more courageous or the culture is more open, but because the architecture of the process is designed to receive what the organisation already knows.
The question is not how to encourage people to speak up. It is how to build a process whose architecture does not require them to.
The cost of not having this
The cost is most visible in the aftermath of failures. But it accumulates long before the failure, in the slow compounding of decisions made without the distributed intelligence the organisation already held. In the projects that required expensive course correction six months in that a well-designed early input process would have caught in week two. In the strategies that were coherent at the level they were made and incoherent at the level they were executed.
The people who already know are in every organisation. The question is whether the process is designed to hear them.