Evidence before hype: how we think about AI in health

It's easy to make bold claims about what AI can do for your health. It's much harder to back them up. As we've built Feelpo, we've tried to hold ourselves to a simple standard: if we can't explain why a piece of guidance is being shown, we don't ship it.

That means grounding features in existing peer-reviewed research wherever possible, being explicit about the limits of what we currently know, and treating our own models as something to keep validating, not something we get to declare finished.

It also means resisting the temptation to overstate what Feelpo can do. We are not a diagnostic tool, and we don't want to sound like one. We're a tool for understanding everyday patterns (sleep, stress, recovery) clearly and honestly.

As our evidence base grows, we'll publish more detail here and in a dedicated evidence library. If you're a researcher or clinician and want to be part of that process, we'd like to hear from you.