In the hustle-bustle of daily life, we are often too drained to book a doctor’s appointment when unwell. Despite warnings against self-diagnosis, many still google for health information, most of which is unreliable without effective filters.
Jivi.ai, a doctor-led AI platform, says it aims to fill this gap by democratising access to quality healthcare globally.
It was founded in January 2024 by Ankur Jain, former chief product officer at fintech firm BharatPe, and GV Sanjay Reddy, chairperson of fund house Reddy Ventures.
Jivi’s conversational platform combines AI-powered diagnostics with human validation from doctors. It creates personalised health reports with details such as symptoms, and the potential diseases and treatment options based on data from millions of medical journals and years of research.
For doctors, Jivi helps analyse patient health data and simplifies workflows to free up their time for consultations.
Jain, CEO of Jivi, lists three problems faced by medical systems globally — doctor unavailability in smaller towns in India and even in countries like the UK, where a primary healthcare appointment can take weeks; shortage of doctors leading to overburdening of existing personnel; and costly treatments.
AI health assistant
In the absence of a healthy doctor-patient ratio, AI platforms can widen access to medical care, Jain says.
“At Jivi, we are building personal health AI assistants for patients and a co-pilot for doctors to help improve the quality of diagnosis,” he says.
The personal AI assistant can analyse the user’s blood test report and, if there are any discrepancies, it can even offer a diagnosis, he says.
“We are not just covering run-of-the-mill symptoms like cough and fever, but also have 18 specialists in our system, including general medicine doctors, cardiologists, dermatologists, radiologists, and ayurvedic doctors. We are going deeper for chronic diseases.”
Explaining the user interface, he says if a patient reports having a fever then the AI assistant will remind them that they also had a fever two weeks ago and will list the antibiotics used back then. It will then try to determine why the fever has resurfaced and whether it is a continuation of the previous episode.
“It’s more of a personal doctor compared to a normal web search,” Jain says.
Health-only info
While the AI model has infinite memory, it can connect applicable dots to provide a diagnosis, he says. For example, a fever experienced six months ago may be irrelevant to the user’s current health status.
The multilingual platform is accessible in English, Hindi, Telugu, Tamil, Kannada, Punjabi, Marathi, and Gujarati, besides Spanish, German, French, Portuguese, and Italian.
“If you ask horizontal platforms like ChatGPT whether, say, an apple is good, it may talk about the fruit, the company, or something else. When you ask Jivi, the focus is on the medical domain, so it’ll give a nutrition- or diet-related answer,” Jain explains.
Jivi’s proprietary AI model, Jivi MedX, recently ranked No. 1 on the Open Medical LLM Leaderboard, overtaking Google’s Med-PaLM 2 and OpenAI’s GPT-4. Jivi uses advanced agentic AI architecture to provide personalised health insights.
Fee model
“For the B2C (business-to-consumer) segment our core product will be free with subscription plans. For the B2B (business-to-business) segment dealing with doctors and hospitals, it’ll be an SaaS (software-as-a-service) product. They will license it,” Jain says.
Computer scientist Andrew Ng-led AI Fund, backed by Sequoia Capital and Softbank Group, has invested an undisclosed amount in Jivi. This marks the fund’s first investment in India.
Jivi is currently beta testing with leading hospitals and is expected to be available for public use early next year.