How Ex-BharatPe CPO’s AI Startup Is Set To Change The Face Of Indian Healthcare

The Indian healthcare sector is ailing and you may ask why. Well, the country has a doctor-to-patient ratio of merely 0.7 doctors per 1,000 people, abysmally lower than the World Health Organisation’s (WHO) average of 2.5 doctors per 1,000 people. 

Then, there are public hospitals which, albeit affordable, are understaffed, poorly equipped and largely limited to urban areas. 

Turning to private avenues for treatment entails out-of-pocket expenses, which could push a family into poverty. As per a NITI Aayog report, 7% of India’s population, nearly 10 Cr people, are pushed into poverty every year due to the money they spend on healthcare. 

It was the very equation that haunted Ankur Jain, the former chief product officer (CPO) of fintech unicorn BharatPe, quite often. Born in a family of doctors, Jain couldn’t have understood it better, triggering him to do something about the weak state of the Indian healthcare system. 

Notably, for him, his second entrepreneurial opportunity came knocking when GenAI mania began gripping the world. Speaking with Inc42, Jain said that patients in India often complain about not being able to get the best treatment while healthcare professionals are often overburdened. 

To solve these challenges, Jain partnered with venture capital (VC) firm Reddy Ventures’ chairman Sanjay GV Reddy to float his second venture Jivi. 

Founded in December 2023, Jivi has built an advanced AI conversational platform that can help patients independently diagnose their illness based on their symptoms. 

As per the company, its platform enables users to get a basic healthcare screening for several chronic diseases such as diabetes, migraine, and tuberculosis, among others. Besides, customers can check their heart rates and their “emotional well-being”. Jivi also claims to have partnered with Tata 1mg for telemedicine and lab testing. 

The platform went live in April this year after operating in stealth mode for five months. It also released its Android and iOS apps in beta mode in September 2024. 

The Genesis Of Jivi AI

Jivi is not Jain’s first tryst with healthcare. Right after receiving his degree from Stanford University in the US in 2006, he worked with a US-based early stage health search engine, Kosmix. 

Speaking with Inc42, Jain credits the experience gained while building Kosmix as the base for founding Jivi. 

His AI experience comes from his stint at Walmart Labs (where he built advanced AI and analytics platforms) and an AI-powered travel tech startup called Instalocate, which he founded in 2017 and shut down in 2020 due to the Covid-19 pandemic. 

Thereafter, he joined BharatPe as the chief product and technology officer in 2020 and was later elevated to the position of CPO. Yet, he yearned to build something that had a far-reaching impact on people’s lives globally. 

He saw AI changing the face of Indian healthcare, and hence the idea of Jivi took birth. While emphasising that the idea of integrating AI or machine learning (ML) with healthcare has been around for several years, the impact of such AI and ML-led tools in healthcare was limited due to their cost-heavy nature. But this was all set to change, at least what Jain was seemingly betting on back then.

“Things took a different turn when research papers came out on the use of GenAI in healthcare, which could allow us to build AI solutions that could drive real-world solutions for healthcare. Since 2018, I have been researching to understand the ways to build AI solutions that could have a lasting impact,” Jain said.

A Sneak Peek Into Jivi’s Tech Stack

The research led Jain to the understanding that such a healthcare solution would require two supporting AI elements – transformers that can translate languages and summarise texts and large language models (LLMs) that can sift through images and voice commands.

Subsequently, Jain developed an in-house proprietary AI model Jivi MedX, which was built atop existing open-source LLMs and trained using an extensive exclusive medical dataset, including millions of medical research publications, journals, clinical notes, and other sources. 

Thereafter, Jain sat down to address one of the key problems facing AI models – hallucinations. The matter takes a serious turn in the case of healthcare as hallucinations could lead to chatbots throwing up random diagnoses and false positives, which may lead to unnecessary hassle and sometimes expenses.

To address this problem, Jivi partnered with multiple doctors and medical institutions to fine-tune its AI model and curb hallucinations. Subsequent scrutiny from top medical professionals in India and approval led the company to finally introduce the beta phase earlier in September. 

Interestingly, Jivi is not limiting itself to just one chatbot offering. The healthtech startup recently announced its second large language model, Jivi-RadX-v1, which is a vision-based LLM that assists radiologists in making faster and more accurate decisions. As per Jain, it deciphers complex image data and delivers sharper insights. 

The Way Forward For Jivi

Even though Jain claims to be “satisfied” with the performance of Jivi’s AI model, Jain has set his eyes on further fine-tuning the LLM. He believes that the timeline for the full launch of Jivi would be early next year. 

Speaking with Inc42, the cofounder claims that Jivi will continue to remain free even after the beta phase. But, how does the founder plan on making money?

Well, the company is eyeing a three-pronged strategy to drive monetisation. It already earns a commission (for each transaction) from user referrals to online marketplaces. As of now, the platform is only live with Tata 1mg. 

The second potential revenue stream would come from licensing Jivi AI software to hospitals, doctors and chains. The third source is building an LLM for insurance companies, which would help them improve the underwriting process. 

Despite the niche offerings, there appears to be no dearth of direct competitors in the AI-led healthcare space. Jivi locks horns with the likes of Parcha, YugasaBot, and Augnito, among others, in the burgeoning segment. 

However, Jain appears to be unfazed by the rising tide of contenders. For him, Jivi has a big ace up its sleeves in the form of its investor and third cofounder, Andrew Ng

Ng is an AI research pioneer, the founder of DeepLearning.AI, chairman and cofounder of Coursera, and an adjunct professor at Stanford University.

For Jain, the support of the AI wizard opens the way to build a deeper tech stack and global insights necessary for a global-scale expansion. 

While Jivi’s ambitions are lofty, it remains to be seen how it walks a tightrope between regulation and innovation as healthcare is one of the tightly governed spaces in the country. However, for now, Jivi’s team has a more equitable, precise and personalised healthcare-focussed AI stack to build for the masses. 

[Edited By Shishir Parasher]

Note: We at Inc42 take our ethics very seriously. More information about it can be found here.