Blog

AI-Assisted Dose Logging

Unfair Team • January 23, 2026

The reason most people stop logging supplements is not laziness. It is friction. Every extra tap, every field you have to fill in manually, every moment you spend trying to remember what you took at 7 AM when you are logging at 10 PM adds up. Within two weeks, most manual tracking habits break down. AI-assisted logging exists to solve that specific problem: keeping the data flowing even when your day does not cooperate.

The friction problem

Traditional supplement logging asks you to do too much at the wrong time. You have to remember which supplements you took, recall the exact dose, note the time, and ideally add a response label ("good energy," "slept poorly"). That works on day one. By day ten, entries become sparse, backdated, and vague.

The behavioral science is straightforward. Habit formation research shows that reducing the number of steps in a behavior increases the likelihood of repetition. ¹ AI logging applies this principle directly:

• Prefilled entries based on your routine. If you take creatine 5g, omega-3, and vitamin D every morning, the AI presents that stack as a single confirmable entry. One tap instead of three manual inputs.
• Timing inference. The system learns that you typically dose your morning stack between 7:00 and 8:30 AM and your evening stack around 9 PM. Prompts arrive at those windows, not at arbitrary times.
• Structured response labels instead of free text. Instead of writing "felt pretty good today I think," you select from labels like "sharp focus," "low energy," "good sleep," or "GI discomfort." These labels are consistent across days, making trend analysis possible.
• Skipped-dose logging. Missed doses are data too. AI prompts you to confirm a skip rather than leaving a silent gap in the record, which matters when you are comparing adherence rates across weeks.

What a full day of AI-assisted logging looks like

Here is a concrete example for someone running a cognitive performance stack alongside a foundation protocol.

Morning (7:30 AM prompt)

The AI presents your typical morning stack:

You confirm all four with a single tap. Total time: under 5 seconds.

Midday (12:30 PM prompt)

The AI asks for a brief cognitive response note using structured labels:

> How is your focus today? Sharp / Normal / Scattered / Fatigued

You tap "Sharp." The label is timestamped and linked to this morning's stack. Total time: 2 seconds.

Pre-workout (4:00 PM prompt)

Your training-day stack appears:

You confirm. Total time: 3 seconds.

Evening (9:30 PM prompt)

Your sleep support entry:

You confirm and add a structured sleep note when prompted the next morning:

> Sleep quality last night? Deep / Normal / Light / Disrupted

This daily pattern takes under 30 seconds total and produces a complete, structured log that is immediately usable for weekly review.

Why structured labels matter more than detailed notes

Free-text logging feels thorough but produces data that is nearly impossible to analyze. "Felt kind of tired but also had a bad night" is a human thought, not a searchable data point. Structured labels solve this by constraining your input to categories you defined at the start of the protocol.

Good label sets for common goals:

When you use the same labels every day, patterns emerge fast. If "Scattered" shows up on days you skipped caffeine + L-theanine, and "Sharp" shows up on days you took it, you have a signal worth investigating. Free text would never surface that pattern without manual re-reading.

What AI logging does not do

It is worth being honest about limitations:

• AI does not verify that you actually took the supplement. A confirmed log is a self-report, not a blood test. This is no different from manual logging, but it is worth remembering that adherence data is only as honest as the person tapping "confirm."
• AI does not replace your judgment. Prefilled stacks are suggestions based on your history, not prescriptions. If you changed your protocol yesterday, you need to update the template. The AI adapts over time, but the first few days after a change require manual edits.
• AI does not establish causation. A correlation between "Sharp focus" labels and your morning stack does not prove the stack caused the focus. Confounders (sleep quality, caffeine timing, workload) still matter. AI logging makes it easier to collect the data you need for a structured review, but the review itself still requires your thinking.

Privacy and data control

Supplement logs contain personal health data. A responsible AI logging system should:

• Store data locally on your device unless you explicitly choose cloud sync. Your supplement history, response labels, and health notes should not be training data for someone else's model.
• Make all AI suggestions editable. Every prefilled entry should be changeable or deletable. You own the log.
• Show the reasoning behind suggestions. If the AI recommends logging a supplement you did not plan to take, you should be able to see why (e.g., "You took this 6 of the last 7 days").

AI logging in Unfair

Unfair's logging workflow is built around the principles above. Morning, midday, and evening prompts arrive at your actual dose windows. Entries are prefilled with your current stack and confirmable in a single tap. Structured response labels link directly to your active protocol, so weekly reviews show your adherence rate and response trends side by side. Skipped doses are tracked, and every AI suggestion is editable. Data stays on your device.

The goal is not perfect logs. The goal is logs that are consistent enough to support real decisions about what to keep, what to adjust, and what to remove.

Continue with Supplement Tracking Best Practices, Best iOS Apps for Supplement Tracking, and The Role of AI in Supplement Recommendations.

References