Truth be told despite the view that History has a habit of repeating itself, it does not. However this does not preclude us from drawing lessons from it. In the 18th and 19th centuries, landowners enclosed common land, transforming shared resources into private estates. Wealth accumulated to those who controlled the land. Everyone else paid for access.
Today, something remarkably similar may be happening with intelligence itself.
Satya Nadella (CEO, Microsoft) recently warned that organisations that fail to train AI on their own operations risk ceding value to a handful of general purpose models. It was a polite way of describing what I see as ‘The Great AI Enclosure’.
For decades, businesses invested fortunes building proprietary knowledge. Industry expertise. Customer insight. Operating procedures. Risk models. Decision frameworks. The accumulated wisdom that differentiated them from competitors. Then AI arrived and, in a fit of collective excitement, many organisations began connecting that intellectual capital directly into frontier models owned by somebody else.
The promise was irresistible; faster decisions, greater productivity and lower costs. The danger is less obvious. If the model increasingly becomes the mechanism through which decisions are made, recommendations generated and work performed, where exactly does the value accumulate? Not necessarily with the organisation itself.
The uncomfortable possibility is that many enterprises are helping build dependency on intelligence platforms that may ultimately commoditise the very expertise that once differentiated them, and the great perversion is they will likely be paying premium ‘tokenised’ rates for the luxury of doing so!
I see a missing layer here. Currently much of the market discussion focuses on AI harnesses; the frameworks that orchestrate models, agents, prompts, tools, memory and workflows. These are important but a harness without organisational intelligence is simply a more sophisticated way of consuming somebody else’s thinking.
The real strategic asset is the Organisational Intelligence Layer, a substrate that is the codified representation of how an organisation thinks, decides and governs itself. It contains business logic, risk appetite, regulatory interpretation, operational expertise, trust requirements, customer understanding and institutional knowledge. In effect, it becomes the digital embodiment of the organisation itself.
The harness then serves as the execution mechanism through which that intelligence is applied to whichever frontier model happens to be best today. Models become replaceable. Organisational intelligence does not. This I see has implications for the future of digital sovereignty and where the debate is likely to evolve.
As I have written about before, for years, sovereignty discussions focused on where data resides, who owns infrastructure and which jurisdiction governs access.Those questions remain important but they increasingly address yesterday’s problem.
The defining sovereignty question of the AI era is becoming, Who conditions the intelligence that thinks, decides and acts on behalf of the organisation?
An enterprise may host its data locally, operate sovereign infrastructure and use a domestically regulated cloud provider. Yet if its decision making, risk interpretation and operational logic are increasingly conditioned by external intelligence platforms, how sovereign is it really?
The practical manifestation of digital sovereignty therefore becomes not simply infrastructure ownership. Ownership of the Organisational Intelligence Layer.
The organisations that thrive in the AI era will not necessarily build the largest models. They will build the strongest intelligence layers and use harnesses to project that intelligence across an evolving ecosystem of models, agents and automation.
In an echo of the first industrial revolution that enclosed land. The AI revolution may enclose intelligence. The winners will be those that own the estate, the rules that govern it and the harness that allows them to cultivate it.
My warning to organisations caught in the AI ‘Gold Rush’ fever is be careful, be very wary not to mistake convenience for sovereignty. Otherwise, you may find yourself paying rent on an intelligence estate built from your own knowledge, expertise and intellectual property.
Leave a Reply