Russian AI Startup ERNA Transforms Heart Diagnostics with ECG Analysis

Artificial intelligence is playing a growing role in medicine, helping doctors tackle complex challenges. One Russian company, ERNA, has developed an AI system to analyse heart data—specifically electrocardiograms (ECGs)—and detect over 30 cardiac conditions. The technology is still being tested and certified, but it signals a shift in how healthcare professionals may soon work alongside AI tools.

ERNA was founded in 2019 as a developer of medical software and diagnostic systems. The company gained recognition in 2021 after winning the Skolkovo Startup Challenge. By 2023, it began building Cardium, a clinical information system designed to streamline diagnostics.

In 2024, ERNA completed its first AI model for ECG analysis. The system was trained using anonymised patient data from multiple Russian regions. It can now identify a wide range of heart pathologies, from arrhythmias to structural abnormalities. The model is currently undergoing certification as an official medical device, though no public data yet confirms how many hospitals have adopted it.

Beyond ECG analysis, AI's broader potential in medicine includes sorting diagnostic data, flagging urgent cases, and acting as a second opinion. However, integrating such tools into daily practice brings challenges. Doctors may ignore AI alerts if the system generates too many false warnings or disrupts workflows. The real test lies in creating a platform that is secure, intuitive, and genuinely useful—one that helps clinicians save time and lives without adding unnecessary complexity.

For physicians, the rise of AI means learning new skills. Rather than replacing judgement, the technology demands that doctors become AI managers—verifying conclusions, spotting limitations, and making the final call on treatment.

ERNA's AI model for ECG analysis is still in its early stages, with certification pending and adoption rates unclear. If successful, it could change how heart conditions are diagnosed and managed. The bigger question remains: can AI systems be designed in a way that doctors trust, rely on, and use effectively in real-world medical settings?