Enterprise AI investment is substantial and growing. The tools are genuinely valuable: they process large volumes of documented knowledge, find patterns that human analysts would miss, generate drafts and summaries at scale, and surface precedents from the full body of what an organisation has documented. For everything that works on existing documented knowledge, AI is transformatively capable.
The question most organisations are now asking — after the initial deployment, after the dashboards, after the pilot programmes — is where the limit is. Not the technical limit. The epistemic limit. The point at which AI reaches the boundary of what it can produce, not because of insufficient compute or training data, but because the knowledge that would change the outcome does not yet exist in documented form.
The layer AI cannot generate
AI systems are sophisticated archives of what has been expressed. The intelligence in them reflects the intelligence in the documents that trained them. For questions that can be answered by recombining existing documented knowledge — summarising, analysing, generating options within known parameters — this is exactly what is needed.
For questions that require understanding that has never been expressed — the strategic tension that exists in the organisation but has not been surfaced, the context that the customer has never been asked to articulate, the perspective from the person closest to the problem who has never been in the room where the decision was made — AI cannot help. Not because it is not powerful enough. Because the knowledge it would need does not exist in any form it can access.
AI can find anything in what has already been expressed. Collective sensemaking builds what has not yet been expressed — and makes it available to be worked with.
The source layer
Hunome is not a competitor to AI tools. It is the layer that produces what AI tools cannot: genuinely new human understanding, structured, characterised, and preserved so that it can be worked with.
A SparkMap on a strategic challenge is not a document that AI will summarise. It is a deliberative process through which the people closest to the question build shared understanding — with the reasoning preserved, the epistemic diversity held structurally, and the connections between ways of knowing made visible. When that understanding is built, it exists in a form that AI tools can work with. But it was not there before. The organisation's AI stack was not working with it, because it did not exist yet.
The sequence matters: deliberation first, analysis second. Not because analysis is less valuable — it is genuinely valuable. But because the analysis of undocumented understanding requires the understanding to be documented first. And that documentation is the output of structured collective deliberation.
What this changes for enterprise AI programmes
Most enterprise AI programmes are built around a documented knowledge layer that was created for a different purpose — the accumulated records of transactions, communications, and analyses that exist because operations require them. This layer has limits: it reflects what was considered important enough to record, shaped by the systems that recorded it.
Collective sensemaking creates a new documented layer: structured human deliberation, characterised by epistemic type, preserved as a living map of how the organisation understands its challenges. This is knowledge created specifically to be worked with — not a by-product of operations, but a deliberate product of collective understanding-building.
AI tools that work on this layer work on richer material: not just what happened, but how the organisation understood why it happened, what it means, and what the people closest to it believe about what should come next.
What the two together do
AI processes the layer that exists. Collective sensemaking builds the layer that needs to exist. Together, they close the loop that neither closes alone: the documented knowledge layer expands as the organisation builds understanding that has never been expressed; the AI tools that work on that layer produce more relevant outputs because the inputs are richer; and the collective sensemaking process becomes more focused because the AI layer can surface patterns in the documented knowledge that inform where deliberation should go next.
Neither replaces the other. Each enables the other to do what it does better.