Treating AI's output like a weather report

TITLE: Your client's lived experience is not an error term BRIEF: Medical AI promises precision but delivers a blurry snapshot. Your job as a coach is to bring the client's lived reality into focus, treating the data as a single instrument, not the whole orchestra.

The latest pitch is that AI will give you and your clients an objective, data-driven path to health. A clean dashboard, a clear score, an unbiased truth. The problem is, your client feels terrible, and their wearable says their sleep score is 92. The data has become another stick to beat themselves with, another way to feel like they are failing at being well. This is not a client problem. It is a category error. The AI is a powerful tool for measuring a specific, isolated thing. It is a terrible tool for measuring a human life.

Your client comes to you confused. They’ve done all the things. Their HRV is up, their resting heart rate is down, yet the dread is still there. The exhaustion is bone-deep. The AI’s report is a weather forecast for a different city. It’s technically accurate, but useless for deciding if you need a coat right here, right now. As a coach, your job is not to debug the data. It is to hold the data alongside the human, to treat their lived experience not as an anomaly or an error term, but as the primary signal. The AI is giving you a note; the client is giving you the music.

The Data is Exteroceptive, The Feeling is Interoceptive

Your client’s wearable is an expert at exteroception. It measures things from the outside: heart rate, skin temperature, movement. It collects externalised data points and runs them through an algorithm. The output is a neat, tidy graph. This is useful, but it is profoundly incomplete.

The client’s felt sense—the tightness in their chest, the buzz of anxiety, the heavy fog of fatigue—is generated by interoception. This is the nervous system’s sense of the body’s internal state, broadcast from the organs and tissues to the brain. Think of it as the body’s internal radio, constantly streaming updates on its operational budget. Is there enough fuel? Is the engine overheating? Is it safe to downshift?

The AI hears the engine RPM. It doesn’t know if the car is speeding downhill in neutral or climbing a steep mountain in first gear. The interoceptive signal is what tells you about the strain on the engine, the smell of burning oil, the feeling of the wheels losing grip. When a client says, "I feel wired and exhausted," and their wearable says, "You had 8 hours of restorative sleep," the AI is not wrong. It just isn't listening to the right station. It’s reporting on the map, while the client is describing the territory.

Related anchors: sleep anchor · HRV anchor · skin anchor

The Context Collapse of an Algorithm

An algorithm has no sense of context. It sees a spike in cortisol and labels it ‘stress’. It cannot distinguish between the stress of a joyful surprise party and the stress of a sudden redundancy. It cannot tell the difference between the sympathetic activation of a good workout and the sympathetic activation of reliving a traumatic event. It’s all just data.

This is the algorithm’s fatal flaw: context collapse. A high-stress reading on a Tuesday in July means something entirely different for a programmer in Silicon Valley bracing for layoffs with no PTO in sight, versus a marketing manager in Madrid the week before the entire country shuts down for August. The physiology might look similar for a moment, but the narrative, the resources, and the recovery path are worlds apart.

As a coach, you are the context engine. You are the one who knows that your client’s ‘good’ sleep score came after a day of dissociating through back-to-back meetings. You know their ‘low stress’ day was actually a day of numb, frozen shutdown. The data point is the beginning of the inquiry, not the end. The Kokorology framework teaches coaches to use these data points as prompts for curiosity, not judgement. You are not there to validate the machine; you are there to validate the human.

Common Questions

How do I explain this gap to a data-obsessed client?

Treat the wearable's output like a single, slightly eccentric instrument in an orchestra. It might be playing the right notes, but it doesn't know the tempo or the key of the whole piece. Your job, with them, is to listen to the whole orchestra—their energy, their mood, their digestion, their felt sense—to understand the music. The data isn't wrong, it's just a soloist.

What if the AI data directly contradicts what my client is reporting?

Believe the client. Their subjective, interoceptive experience is the ground truth of their system's state. The data’s contradiction is the most interesting part of the puzzle. Ask: "The watch says you were calm, but you felt anxious. What was happening for you then? Where in your body did you feel that anxiety?" This turns the contradiction from a problem into a doorway to deeper self-awareness.

Is there a simple way to integrate wearable data without letting it take over?

Use it as a bookend, not the main event. Look at the data together at the start of a session for maybe two minutes to spot a pattern worth discussing. Then, put it away. The rest of the session is about the client’s experience. This teaches them to privilege their felt sense over the device’s report card.

Should I recommend my clients stop using their wearables?

Not necessarily. For some, it can be a useful tool for connecting actions to outcomes. The goal isn't to reject the data, but to right-size its importance. The reframe is to treat it as a curious observer, not an authoritative judge. If it's consistently causing more anxiety than insight, then suggesting a break can be a powerful intervention.

Closing

• Ready to build a practice that privileges the human over the algorithm? The Kokorology Coach Certification gives you the frameworks to work with the nervous system in a way that data alone never can.
• Need to get your hands on this approach right now? The Kokorology Toolkit contains the core protocols for nervous-system-led coaching, including how to work with client data without getting lost in it.
• Start here for free: Download the guide on A Nervous-System Approach to Client Goal-Setting, and learn how to move beyond metrics and into meaningful change.

TL;DR

Medical AI and consumer wearables provide exteroceptive data—objective but context-free snapshots of physiology. This often clashes with a client's interoceptive reality, their subjective felt sense. This isn't a failure of the client or the tech; it's a context collapse. The coach's role is to be the context engine, treating the AI's output as a single data point, not the definitive truth. Your job is to privilege the client's lived experience, using data contradictions as entry points for deeper inquiry, not as errors to be corrected.

Sources

• Adam, E. K., et al. (2017). Diurnal cortisol slopes and mental and physical health outcomes: A systematic review and meta-analysis. Psychoneuroendocrinology, 83, 25–41.
• Park, Y. et al. (2024). Many medical-AI models fail in the real world. Nature, 625, 438–439.
• Shorer, S., et al. (2021). The role of interoception in the regulation of emotions: A pilot study. Frontiers in Psychology, 12, 651994.
• Sonuga-Barke, E. J. S., et al. (2024). Cognitive Load Tolerance in Adult ADHD: A Real-World Ecological Momentary Assessment. Molecular Psychiatry.
• Germeroth, L. J., et al. (2024). Consumer-grade wearable device data for clinical use. The Lancet Digital Health, 6(1), e4-e5.