75% of Healthcare AI Startups Fail Clinical Validation, Bengaluru Conclave Told

Nearly 75% of healthcare AI startups never make it into doctors’ cabins because they fail clinical validation, according to Dr Harsh Atul Hirani, clinician and co-founder of DocYantra, speaking at the Healthcare and Artificial Intelligence (HAI) Conclave 2026 in Bengaluru on Friday.

Hirani demonstrated AI-powered clinical agents capable of analysing 24 hours of patient data and generating dosage recommendations in under 30 seconds, compared with the four to six minutes typically required by clinicians. ‘AI does not create better doctors, but better doctors use AI to create a greater impact,’ he said, adding that such systems must operate within legal frameworks and safeguard patient confidentiality.

The conclave, organised by the Global Healthcare Academy, also featured Dr Shravan Subramanyam, managing director and group CEO of BPL Medical Technologies, who described how AI could build a ‘digital twin’ of a person’s body to monitor health and predict disease risk from before birth through old age. He noted average life expectancy has risen from the mid-30s to the 70s as care shifts from reactive treatment to continuous, personalised monitoring, while stressing, ‘We are all the human in the loop.’

A lighter counterpoint came from Muralidhar Somisetty, founder and CEO of Wellnesys Technologies, who showcased Yogifi, a Rs 12,000 AI-enabled smart yoga mat with sensor fabric that tracks balance and posture in real time. Somisetty said he built the product after his own recovery from chronic neck and back pain that required spine surgery.

Organisers rounded out the day with demonstrations of AI-powered patient simulators for communication training, framing the event as a snapshot of how India’s health-tech sector is trying to move AI from research labs into everyday clinical practice.