A Belfast startup has won US clearance for an AI layer that rescues the readings smartwatches give up on.
PulseAI said on November 5 that the FDA granted 510(k) clearance for its arrhythmia analysis model, built to sit on top of ECG recordings from wearables and return a structured interpretation for clinical review. Recordings go to the cloud; results come back to the connected app.
The clearance rests on a prospective study of 296 participants published in Heart Rhythm. In that trial the model raised sensitivity for atrial fibrillation and cut inconclusive interpretations by 92% compared with the Apple Watch’s own algorithm. That gap matters because inconclusive traces trigger manual overreads, repeat recordings and follow-up calls.
Smartwatch volume explains the pressure. IDC counted 164 million units sold worldwide in 2025, and most of them record a heart rhythm. Each unresolved trace lands on a clinician.
“Wearables have made ECG recording widely accessible,” chief executive Alan Kennedy said. “Our ambition is to match expanded access with improved accuracy.”
The company, spun out of Ulster University research, is now chasing integration deals with device makers and software platforms rather than building its own hardware. Innovation Ulster Limited and the Mortara Group led its earlier round, with Justin Mortara, former chief of Mortara Instruments, named board chair.
Atrial fibrillation remains the most common sustained rhythm disorder and carries roughly a fivefold higher ischemic stroke risk when untreated, which is why a cleaner read on a consumer trace has clinical weight.
