WHOOP’s GPT-4 Coach Knows Your Body Better Than You Do

The quantified-self movement just got its most conversational upgrade yet. WHOOP, the screenless wearable that tracks sleep, strain, and recovery around the clock, has launched WHOOP Coach — a generative AI feature powered by OpenAI’s GPT-4 that lets members interrogate their own biometric data in plain English. Think of it as a search engine for your physiology, only it knows your lowest resting heart rate and exactly how much sleep debt you’re carrying.

The engineering team fine-tuned GPT-4 on anonymized member data and proprietary WHOOP algorithms, pushing the model far beyond generic fitness advice. Ask Coach for a weekly workout schedule calibrated to a marathon goal, and it doesn’t just spit out a template — it crunches your historical recovery scores and heart rate variability to tell you when to push and when to back off. Jaime Waydo, WHOOP’s CTO, described Coach as something that’s “behind the scenes, looking at thousands of your own unique data points” to surface the most actionable next step.

Early usage data suggests members aren’t just kicking the tires. Forty percent of all questions fielded by Coach are requests for recommendations rather than simple lookups, and the top query is refreshingly earnest: “How can I improve my sleep quality?” Four of the five most popular questions fall squarely in the self-improvement category, signaling that people treat the tool less like a novelty and more like a 3 a.m. personal trainer who never gets annoyed.

Will Ahmed, WHOOP’s founder and CEO, acknowledged the AI hype cycle head-on, arguing that Coach actually cashes the check that the industry has been writing. There’s a real question lurking beneath the launch, though: when your wearable starts giving you advice that sounds unnervingly confident, do you trust it enough to change your behavior? WHOOP is betting yes, and if those recommendation stats hold, the answer might already be written in the data.