Loading...

elth.ai looks to disrupt med-tech market with AI

Loading...

elth.ai, owned by Bengaluru-based HCFAB Health Technologies Pvt Ltd, is an artificial intelligence-based med-tech startup that aims to solve health problem discovery and the absence of a 24×7 medical adviser. Founded by Srinath Akula, an IIM-Indore alumnus who worked with Flipkart earlier, and dental implantologist Akhila Srinath, it is attempting to bridge the communication gap between patients and doctors.

Akula believes 2017 is going to be the year when AI will take over and chatbots will replace apps, notwithstanding the fact that executing AI is complex.

Loading...

"The interest in health-tech, and the number of startups that have come up in this space, is tremendous. Digitisation of records and transactions, laboratory services, sample collection and at-home services, all these have been really great. But as of now many startups are shying away from using AI because of the complexity involved," Akula adds.

Switch to chatbot

During her two-and-a-half-year-long practice, Akhila realised that with most doctors pressed for time, patients are often left with a lot of unresolved queries. "So, we thought there should be somebody sitting between the doctor and the patient to solve the queries of the patients as well as to communicate seamlessly with doctors," she says.

Loading...

When elth.ai was launched in January 2016, it had Android and iOS apps catering to dental problems. However, the founders soon realised that people don't want to keep medical apps on their phones because of space constraints, and delete it immediately afterwards. Hence, they pivoted to a messenger-based platform and launched their first chatbot in September 2016.

"In fact, most of the post-treatment queries can be automated. That's what we're training our bot to understand and answer," Akhila adds.

What it does

Loading...

elth.ai can understand the user's natural language query and suggest solutions ranging from home remedies, preventive care and tests to diet plan and even the right specialist. The chatbot asks questions about symptoms and previous complications and, using these data points, it tries to understand the problem.

Business model

Loading...

The company makes money by charging hospitals for lead-generation. Because of the domain expertise gained through AI, it also provides software-as-a-service (SaaS) applications to big hospitals. On the consumer sider, however, it is a free service.

Funding

Healthcare-focussed startup incubator HealthStart had announced the third batch of its accelerator programme in August last year. Of the 200 companies that had applied for the programme, four got selected and elth.ai was one of them. The startup is now looking to raise pre-Series A funding.

Loading...

USP and competition

In terms of lead generation, the startup faces competition from market biggies Practo and Lybrate.

"The reason why we are superior is AI. Our AI precisely understands what a doctor's strong areas are. And because we have a lot of information about the patient, doctors can make an informed choice. Our USP is that we understand the users much deeper using AI," says Akula.

Loading...

"Also, the patients would prefer quick solutions, rather than wait for 4-6 hours or two days. Our bot gives instant answers to users in terms of what their potential problem could be, suggesting hospitals and doctors," he adds.

Traction

The startup claims to have 10,000 users from more than 45 countries, of which around 80% are from India and 15% from the US. On a normal day, it gets 400-500 users and its month-on-month growth rate has been 300-400%.

The company says that of the 400-500 users who interact every day, 60% are the repeat users.

Challenges 

elth.ai feels India's low internet penetration is not really a challenge, since there are over 1.2 billion users of Facebook messenger who can easily access the bot. Besides, Facebook has given indications that it will make bots available on WhatsApp as well, which has more than 1.5 billion users.

"There will always be a gap, where some user queries won't be understandable, but we're using machine learning to improve. This is a challenge because natural language processing (NLP) can never be perfect. And it's available only in English as of now, because NLP support for other languages is very poor," he signs off.


Sign up for Newsletter

Select your Newsletter frequency