Lotus Health nabs $35M for AI doctor that sees patients for free | TechCrunch
Briefly

Lotus Health nabs $35M for AI doctor that sees patients for free | TechCrunch
"In essence, Lotus is building an AI doctor that functions like a real medical practice, equipped with a license to operate in all 50 states, malpractice insurance, HIPAA-compliant systems, and full access to patient records. The key difference is that the majority of the work is done by AI, which is trained to ask the same questions a doctor would."
"Since AI models are also prone to hallucinations, the company always has board-certified human doctors from top health institutions such as Stanford, Harvard, and UCSF review the final diagnoses, lab orders and medical prescriptions. Lotus has developed an AI model that, similar to OpenEvidence, synthesizes the latest evidence-based research with a patient's history and clinical answers to generate a treatment plan. "AI is giving the advice, but the real doctors are actually signing off on it," Dhaliwal told TechCrunch."
"KJ Dhaliwal (pictured left), who in 2019 sold the South Asian dating app Dil Mil for $50 million, says he has been thinking about the inefficiencies of the U.S. healthcare system ever since he was a child acting as a medical translator for his parents, and he saw the advent of LLMs as an opportunity to do something about it."
KJ Dhaliwal launched Lotus Health AI in May 2024 as a free primary care service available 24/7 in 50 languages. Lotus raised $35 million in a Series A round, bringing total funding to $41 million. The platform combines large language models with licensed clinical infrastructure, malpractice insurance, HIPAA-compliant systems, and full access to patient records. AI performs the bulk of clinical questioning and drafts diagnoses, prescriptions, lab orders, and referrals. Board-certified human physicians from top institutions review and sign off on final medical decisions. Urgent or emergency cases are directed to in-person urgent care or emergency departments.
Read at TechCrunch
Unable to calculate read time
[
|
]