Viviano Cantu is the AI Product Lead at WHOOP, but his path there did not start in product. He joined WHOOP in 2021 as a software developer, led the storefront team, and then in March 2023 moved over to start a new team around coaching. That same month, he and a few colleagues started ordering pizza on Friday afternoons and hacking on whether an LLM could actually coach someone throughout their day. Six months later, WHOOP Coach launched as the first generative AI coach in a wearable. He has led AI at WHOOP ever since, and along the way his team built AI Studio, an internal platform that lets anyone at WHOOP go from idea to a working AI agent in under ten minutes. In this episode of Product Talk, host and Planned Parenthood Direct Chief Product Officer Hannah Park sits down with Viviano to dig into what separates a weekend health chatbot from a coaching product that reasons over one person’s physiology consistently for years, how WHOOP thinks about evals, synthetic personas, and the my memory feature, and what his engineering background actually changed when he crossed over into product.
In this episode, we cover:
(1:56) The Boston Rat Czar side project, and what it says about how Viviano approaches problems
(3:20) How Whoop Coach started: Friday hack sessions, GPT-3, pizza, and a prototype called the Whoop Bot
(8:33) The GPT-4 pivot mid-build, and why pushing the timeline was the right call
(10:41) The gap between a weekend AI demo and a coaching product that knows one person over years
(11:56) How WHOOP uses synthetic personas to run evals without touching real member data
(13:35) Who decides what a good coaching answer looks like, and how AI Studio made that process faster
(17:25) What building AI Studio did to how the team thinks about prioritization
(20:00) Where WHOOP intentionally put the friction, and why PII never touches the model
(21:39) The new role being developed at WHOOP that sits between engineer and PM
(25:02) What entering the clinical space with EHR sync taught Viviano about precision versus recall
(27:43) Why WHOOP lets members see and edit everything the AI knows about them
(30:34) What Viviano brought from engineering to product, and what did not survive the move
(33:01) The tweet about AI-assisted coding discipline, and what he was watching on his team
(34:56) An AI workflow tip for PMs: one folder, parallel threads, and auto-approve
Blog and detailed workflow walkthroughs from this episode: https://productsthatcount.com/wearable-data-into-an-ai-coach/
