I was a girl the first time I visited Ngamatea Station. It is one of New Zealand’s great high-country stations: 36,000 hectares of sheep and beef country in the central North Island, running along the Taihape-Napier Road. I was there with a group of friends, visiting the daughter of the owner, Kate Apatu. I remember the scale of the land, the particular quality of the light in that part of the country. But what I remember most clearly, forty years later, is the homestead.
There was something about that building that stopped you. Not just its grandeur, though it had that. Something harder to name. It felt right in a way that very few buildings do, as though it had always been there and the landscape had simply waited for it. I remember running my hands across a kauri internal door, nearly three metres wide, solid and warm, that led into the media room, and feeling that everything about the place was in some kind of accord. The materials, the proportions, the way it sat on the land.
I found out later how it had come to be. Before the architect drew a single line, he had spent six weeks living in a tent on the hill where the homestead would eventually stand. Six weeks of sitting with the landscape, walking it, feeling the flow of the land, watching where the light moved, attending to what the place itself seemed to want to become. Only after that did the framework begin.
The building that resulted was not just well-designed. It was the physical realisation of something that had first existed in a pre-symbolic space, in the attending and the sensing before the drafting. The architect used every structural skill he had. But those skills served something that had already formed before a single technical decision was made.
That is the distinction this article is about. And it is the distinction that gets lost every time someone describes AI as intuitive.
Why the Word “Intuitive” Has Been Borrowed
AI is increasingly described as intuitive. The outputs feel immediate. They feel relevant. They feel, at times, uncannily attuned to the specific situation you bring to them. This has led to a growing assumption that something like sensing is happening inside the system, that AI is not just processing information but perceiving it.
That assumption is not accurate. And for practitioners working in the intuitive space, the confusion is not merely semantic. It shapes how they understand their own practice, what they think AI can and cannot replace, and crucially, what they continue to develop in themselves and what they quietly stop attending to.
The word “intuitive” has been borrowed from human cognition and applied to machine outputs because the outputs produce a similar effect in the person receiving them. But the effect and the mechanism are entirely different things. Understanding that difference is not a philosophical exercise. It is a practical necessity for anyone whose work depends on genuine intuitive capacity.
The previous article in this series, What AI Can’t Access, established what AI actually is: a system that operates entirely within encoded, symbolic information, predicting probable linguistic sequences from vast training data. It does not access the pre-symbolic stage. It enters the cognitive process after human perception has already done its work. This article takes that foundation and asks a more specific question: why, given that structural reality, do AI outputs so consistently feel intuitive to the people receiving them? And what does that misidentification cost?
Why the Output Feels Intuitive
The illusion is not random. It emerges from specific and overlapping factors, and understanding them matters because they are also the factors that can mislead practitioners about where their own value sits.
The first is compression. AI collapses what would normally be a multi-step reasoning process into a single response. The intermediate steps are not visible. Humans interpret this compression as the hallmark of intuitive knowing, because genuine intuition also produces knowing without visible steps. The mechanism is entirely different. The experience of receiving the output can feel similar.
The second is familiar pattern. The response often carries structures that humans associate with insight: clear framing, contextual alignment, relevant metaphor, confident tone. When these elements appear together, the brain categorises the output as intuitively derived. It has learned, across a lifetime of human interaction, that this quality of response usually signals genuine understanding. AI has been trained on the outputs of that understanding. It replicates the signal without the source.
The third, and perhaps the most significant, is projection. Humans actively participate in the illusion. When a response resonates, the person receiving it supplies meaning from their own internal context. The AI provides the linguistic structure. The human provides the significance. The resonance is real, but it is occurring in the person, not in the system. The system produced something that could fit. The human determined that it fit here, for this situation, in this moment.
This is a co-created effect. It is not a property of the machine.
Taken together, these three factors explain why the misidentification is not careless. It is a natural consequence of how humans interpret intelligence. We are wired to read fluent, contextually appropriate, rapidly produced language as evidence of understanding. AI operates precisely within that interpretive habit. The confusion is understandable. It is also, for practitioners whose work depends on the distinction, consequential.
The Operational Lesson
Some years ago I was brought in to conduct an operational review of a large organisation. The work was complex, involving significant structural analysis across multiple layers of the organisation. I used AI for what it does genuinely well: the architecture of complex information, the organisation of frameworks, the mapping of structural relationships. It was useful. It did what it does, and it did it efficiently.
But over time I noticed something. The more the process became language-focused and framework-driven, the further I moved from what I was actually there to do. My strongest capacity in that kind of work is not structural analysis. It is reading the heartbeat of an organisation: sensing what is moving beneath the surface, perceiving what the data and the frameworks do not yet show, hearing what is not being said in a room as clearly as what is.
The framework was beginning to stymie the build. I had drifted from the pulse of the organisation because the tool I was using had no access to pulse. It had access only to what had been encoded, articulated, structured, expressed. Everything that was still forming, everything that existed in the pre-symbolic space of the organisation’s actual life, was invisible to it.
I had to reorientate. To rebalance what the framework could show me with what could only be sensed. The seen with the as-yet unrealised.
This is not a criticism of the tool. It did what it was built to do. The error was in allowing the tool’s domain to gradually expand into territory it was never designed to reach. When that happens, it is not the tool that loses something. It is the practitioner.
The Architect in the Tent
Return to Ngamatea Station and the architect who spent six weeks on a hill before he drew a line.
What he was doing in that tent was not inefficiency. It was not delay. It was the essential first stage of the work, the stage that made everything that followed coherent. He was attending to what the land itself was communicating, in a language that has no words, only feeling and recognition and the slow accumulation of knowing that precedes form.
No framework could have produced what those six weeks produced. The frameworks came later, and they were necessary. But they served something that had already formed in a space no structural tool can reach.
This is what experienced practitioners in any field understand, often without having explicit language for it. The build requires the framework. But the framework must serve the vision, and the vision comes from somewhere the framework cannot go.
AI is an extraordinarily powerful framework tool. It can organise, synthesise, map, and articulate at a scale and speed no individual can match. In the right position in a workflow, it is genuinely valuable.
But it cannot sit in the tent. It cannot attend to the landscape. It cannot feel the flow of the land before the first line is drawn.
What Gets Lost When the Distinction Blurs
When AI outputs are described as intuitive, something subtle but consequential happens. The boundary between sensing and predicting becomes unclear. Practitioners who are uncertain about their own capacities may begin to measure themselves against a system that is doing something categorically different, and find the comparison disorienting.
More concretely, if what AI produces is understood as a form of intuition, then the practitioner’s own intuitive capacity stops being seen as the thing that needs active development. Why cultivate a faculty that the machine apparently already has?
This is where the misunderstanding becomes costly. Not because AI poses a threat to the practice, but because a misunderstood tool shapes the attention of the person using it. Attention that might otherwise be directed toward deepening perceptual accuracy, toward developing the capacity to sense what has not yet been articulated, goes instead toward the interpretive layer, the layer where AI already operates, and where the competitive pressure is only increasing.
The seeker sitting across from you rarely asks the question they actually came with. What is unspoken is frequently the heart of the matter. What is still forming, still unarticulated, still existing only as a felt sense in the seeker’s body and experience: that is the territory the practitioner must be able to reach.
AI cannot reach it. Not because it lacks the data. Because that territory, by its nature, has not yet become data.
Simulation Is Not Origination
The clearest way to hold this distinction is through a simple contrast.
AI can simulate the expression of intuition. It cannot originate the process of intuition.
Simulation operates within symbolic systems. It recombines existing patterns into outputs that resemble intuitive insight. It does this well enough that the resemblance is often convincing. This is precisely what produces the three-factor illusion described above: the compression, the familiar pattern, the invitation to project meaning onto a structure that has been optimised to receive it.
Origination requires access to the pre-symbolic stage. It requires perception before language, not after it. In tarot, that shows up when cards activate meaning rather than contain it. It requires something to arise from within the practitioner’s own perceptual system, from the accumulated, unencoded, unarchived knowing that lives in the body and the attending and the years of practice.
The homestead at Ngamatea did not arise from a database of architectural decisions. It arose from six weeks of unstructured attending, and from the trained capacity of one practitioner to receive what that attending produced and bring it, carefully, into concrete form.
That sequence, sensing first, framework after, is not a romantic preference. It is the operational structure of all genuinely intuitive work.
Understanding it clearly is what allows a practitioner to use AI well, in the position it belongs, without allowing it to migrate into the stage where it has no business being.
The next article, Pre-Verbal Knowing, examines the science behind what is happening in that pre-symbolic stage: what pre-verbal knowing actually is, what the research shows, and why it is not only trainable but the most important capability a practitioner in an AI-saturated world can develop.
Leigh Spencer is the founder of Tides of Knowing and founder of The COMPASS Method™, a framework for the conditions of attention that make intuitive reading reliable under pressure. With 30 years in professional journalism and 40 years as a tarot reader and intuitive practitioner, she writes at the intersection of symbolic literacy, perceptual development, and the changing landscape of human knowing.