How August AI is helping users access reliable health information

During the pandemic, entrepreneur Anuruddh Mishra was diagnosed with rheumatoid arthritis. “I started picking up some pain in my fingers. That slowly sort of started growing. As I couldn’t really go out, I did what most Indian people would do–I went and got a lab test and sent it to a Delhi-based health platform. On chat, the doctor at the other end told me I have rheumatoid arthritis,” Mishra tells YourStory. 

“However, this was a misdiagnosis. I ended up sort of realising that healthcare is broken in some way for the patient,” he says. 

Recognising the widespread gap in the healthcare system, Mishra started August AI in 2022. It is a health assistant chatbot that offers expert medical insights personalised to every user based on their health profile and history. 

“Several cities in India lack access to quality healthcare... some regions lack even basic medical services,” Mishra explains. 

This disparity highlights the potential for technology to bridge the gap, particularly by providing health information in the local language, he adds.

Accessible round the clock on WhatsApp, the Bengaluru-based startup aims to enhance medical consultations for users with extra support and guidance instead of providing diagnosis. 

The startup has a 11-member team, which comprises of experienced doctors, engineers, and data scientists. 

Mishra says, since inception, over a million users have turned to August AI for health information.

Chat-based WebMD 

Mishra got the idea of August AI as part of Get Beyond Health, a healthtech startup he co-founded in 2022 along with Anand Rao. 

Get Beyond Health was providing personalised health cards for users by gathering extensive medical and healthcare data from various sources to deliver precise evaluation for patients. 

“At that point, we began exploring how to effectively explain health to individuals. We decided to convert lab reports into understandable metrics to simplify the information. After conducting over 1,500 health consultations, we focused on ensuring that our approach was centred around listening and engaging with each person. Our goal was to take as much time as needed to ensure they fully understood their health,” Mishra notes. 

“Language models completely changed the game for us. We had the data and the know-how to communicate health to people. We then developed a clear roadmap and proceeded to build August AI,” he adds. 

However, Mishra says, the pivot from Get Beyond Health to August AI was largely a branding decision. 

“Beyond Health was the original name of the company. The focus was to make healthcare more explainable with the ultimate aim of building a health AI 3-4 years down the line. When LLMs came into the picture, we were caught at the right place, right time, and ended up launching that same thing in six months,” he adds. 

Also Read

Google's AI Detects Diseases from Cough Sounds on Smartphones, Saving Lives

How it works? 

August AI’s chatbot can be accessed by users by scanning the QR code or on their website. Once the chat opens on WhatsApp, users can ask health questions directly.

The assistant is multi-modal–supporting audio, text, and lab reports in PDF format as input, and it provides output in text format.

“It typically takes around a minute to process lab reports. But August AI is a health information platform. It neither gives a diagnosis nor prescribes medication,” says Mishra. 

One of its new functionalities include reading prescriptions and other text-based images. Users will be able to understand their prescriptions, check active ingredients, and verify potential interactions with other medications they’re taking. 

“The reason it’s on WhatsApp is because people already know the controls on WhatsApp. So, they can send a file, they can send a photo, they can text it, or send a voice note. Those are not things we have to teach them in terms of the UX. That did make solving the problem a lot harder,” Mishra says. 

However, accuracy is crucial when it comes to healthcare. With an aim to tackle this challenge, the model uses ensemble refinement—a method where multiple models or reasoning approaches are combined to improve accuracy or prediction outcomes to generate high quality outputs for health-related queries. 

“We started before LLMs (Large Language Models) became mainstream. At that point, saying we’re building a health AI sounded impossible. Taking on models from much bigger companies while bootstrapping with a 4 people team was a major challenge for us,” Mishra says. 

Last year, August AI scored 94.8% on the USMLE (US Medical Licensing Examination), setting the highest score among all AI systems tested on this benchmark. It outperformed notable competitors such as OpenAI’s GPT-4 (87.8%), Google MedPaLM 2 (86.5%), and OpenEvidence (90%), according to a blogpost by Google.

This benchmark of high performance and accuracy blended with the support of images, audio, and PDF format distinguishes the company apart from the likes of other healthcare chatbots such as Mfine and Practo, says the founder. 

Also Read

Meet Myna Bolo, an AI chatbot that answers queries on sexual and reproductive health for women

The market and plans ahead

August AI, which was bootstrapped for the initial two years, secured funding in June this year by Manipal Group Chairman Ranjan Pai’s Claypond Capital in an undisclosed round. 

Mishra says, August isn’t intended to replace a doctor, but it’s worth noting that many people— particularly the younger generation—lack access to a family doctor.

“We see users coming back to August after consulting with their doctor to further clear their doubts or before they go to a doctor to understand what’s happening and who they should be seeing,” the founder explains. 

Besides being a health assistant to general users, the startup follows a B2B model by catering to a major hospital chain operating out of Bengaluru, along with other prominent hospitals, telemed companies, and other institutions in the healthcare industry.  

Apart from India, it currently operates in Southeast Asian markets. 

For now, Mishra’s goal is to make sure people get to the doctor at the right time.

“Making that process simple is crucial. It’s important to have a person’s health data accessible in a way that allows for interactive conversations, where the data is considered as they ask questions. Enhancing this with a more advanced interface is something we believe will be important for us,” he says.