The difference between "I feel bad" and data AI can work with

"I don't feel good" is honest. It's also not something AI can analyze. The gap between how you experience symptoms and what AI needs to find patterns is where most people get stuck — not because they lack information, but because symptoms are messy and the translation into structured data isn't intuitive.

Clinical research on patient history-taking, published in The BMJ, has established that structured symptom reporting significantly improves diagnostic accuracy. The same principle applies to AI: the more structured your description, the more useful the analysis.

You already have the information. It just needs five dimensions.

The five dimensions

Timing is the single most valuable piece of information. "I'm tired" is noise. "I wake up exhausted most mornings and my energy drops again around 2 PM" is data. Research on the Food and Symptom Times (FAST) method, published in PubMed Central, found that precise timing data reveals correlations that retrospective recall consistently misses.

Severity tells AI how much the symptom matters. A twinge is different from pain that changes how you walk. Use a number scale if it helps, but also connect severity to function: can you work through it? Does it stop you? Does it wake you up?

Triggers are the strongest clues about mechanism. Symptoms after eating point toward digestion. Symptoms after stressful meetings point toward stress physiology. Triggers include the obvious (food, exercise, stress) and the subtle (time of day, weather, menstrual cycle phase, sleep quality the night before).

Modifiers — what makes it better or worse — are almost as diagnostic as triggers. Does heat help? Does movement improve it or worsen it? Modifiers help distinguish inflammatory, muscular, metabolic, and functional patterns.

Associated symptoms prevent tunnel vision. Fatigue plus joint pain plus difficulty concentrating is a different investigation than fatigue alone. AI can see multi-symptom patterns if you report the full picture.

An example

Vague: "My stomach has been off. I feel bloated sometimes."

Structured: "For the past week, I've had afternoon bloating after lunch — uncomfortable but not painful, resolves by evening. No diarrhea or constipation. Started after I increased fiber intake last Monday. Happens even with plain food."

The second version gives AI timing, severity, context, what doesn't trigger it, and associated symptoms (none). Now AI can investigate specific mechanisms rather than guessing.

Making this practical

When you're about to ask AI for help with a symptom, spend two minutes on five questions: When does it happen? How bad is it? What was going on? What makes it better or worse? What else is happening at the same time?

The answers are what you feed to AI. The structuring also forces you to notice things you'd otherwise miss — that the fatigue only happens certain days, or that headaches follow specific meals, or that there's a pattern hidden in the noise.

References

Measuring Diet Intake and Gastrointestinal Symptoms in IBS — PubMed Central, 2020. Timing data revealing hidden correlations.
Improving diagnostic accuracy through structured history-taking — BMJ, 2010. Structured reporting and diagnostic outcomes.
The value of patient-reported outcomes — Health and Quality of Life Outcomes, 2015. Structured self-report improving clinical utility.