I’m curious about how WHOOP Coach comes up with its recommendations. What kind of data does it analyze to give me insights?
The WHOOP Coach generates insights by analyzing a range of data collected by WHOOP’s core algorithms, including sleep, strain, recovery, workouts, HRV, and stress. These machine learning algorithms process raw physiological signals to determine things like when you fell asleep, when you woke up, what your heart rate was during a workout, and more.
When you ask WHOOP Coach a question, such as “How did I sleep on June 10th of last year?”, it:
- Identifies the topic and date from your question (e.g., sleep on June 10).
- Pulls relevant data from your WHOOP records for that topic and time period.
- Includes additional context, such as any journal data and behavior impacts that may be relevant.
- Searches for correlations between your question and WHOOP’s Performance Science data and research.
- Removes personally identifiable information before sending it to a large language model (LLM) to generate an answer.
- Adds personal context back in and links to relevant WHOOP app features or articles before presenting the final response.
All of this happens in under three seconds for most questions. WHOOP Coach is designed to provide personalized, fast, and data-backed answers by combining your individual metrics with WHOOP’s models and research.