Based in Bangladesh, Maya is dedicated to making it easier for women to get healthcare, especially for sensitive issues like reproductive and mental health. The startup announced today it has raised $2.2 million in seed funding. The round, which Maya said is the largest raised by a Bangladeshi health tech company so far, was led by early-stage fund Anchorless Bangladesh and The Osiris Group, a private equity firm focused on impact investing in Asian markets.
The funding will be used to introduce new products to Maya’s telehealth platform and expand into more countries. Maya recently launched in Sri Lanka and plans to expand into India, Pakistan, Middle Eastern markets and Indonesia.
Maya uses natural language processing and machine learning technology for its digital assistant, which answers basic health-related questions and decides if users need to be routed to human experts. It has about 10 million unique users and currently counts more than 300 licensed healthcare providers on its platform.
Founder and chief executive officer Ivy Huq Russell, who grew up in Chittagong and Dhaka before moving to the United Kingdom for university, started Maya as a blog with healthcare information in 2011. At the time, Russell worked in finance. She had just given birth to her first child and her mother had recently been diagnosed with breast cancer. Russell told TechCrunch she realized how many challenges there were to seeking medical care in Bangladesh, including financial barriers, a shortage of providers and long travel times to clinics.
She began Maya with the goal of providing trustworthy health information, but quickly realized that the site’s visitors needed more support. Many sent messages through WhatsApp, email or the site’s chat box, including survivors of sexual abuse, rape and domestic violence. After receiving a grant from BRAC, a Bangladeshi non-governmental organization, Maya’s team began developing an app to connect users with medical information and experts.
Maya’s homescreen
“We were very focused on two things,” Russell said. “One is how do we built trust in our community, in their language, because it’s very important that they communicate in the language that they’re comfortable using. At the same time, we realized as soon as we started getting hundreds and hundreds of questions, that we’re not going to be able to scale up if we just have 50 experts on computers typing.”
To support Bengali and regional dialects, Maya spent more than two years focused on developing its natural language processing technology. It collaborated with data scientists and linguists and took part in Google Launchpad’s accelerator program, working on tokenization and training its machine learning algorithms. Now Maya is able to provide automated answers in Bengali to basic questions in 50 topics with about 95% accuracy, Russell said. Out of the four million queries the platform has handled so far, about half were answered by its AI tech.
Many have to do with sexual or reproductive health and the platform has also seen an increase in questions about mental health. These are topics users are often hesitant seeking in-person consultations for.
“Growing up in Bangladesh, we got minimum sexual education. There’s no curriculum at school. Recently in the last one or two years, we’ve also started to see a lot of mental health questions, because I think we’ve made a good drive toward talking about mental health,” said Russell. She added, “it’s quite natural that whatever they couldn’t go and ask a question about very openly in traditional healthcare systems, they come and ask us.”
More consultations are coming from men, too, who now make up about 30% of Maya’s users. Many ask questions about birth control and family planning, or how to support their partners’ medical issues. To protect users’ privacy, consultations are end-to-end encrypted, and experts only see a randomly-generated ID instead of personal information.
In order to understand if someone needs to be routed to a human expert, Maya’s algorithms considers the length, complexity and urgency of queries, based on their tone. For example, if someone types “please, please, please help me,” they automatically get directed to a person. The majority of questions about mental health are also sent to an expert.
Russell said Maya’s approach is to take a holistic approach to physical health and mental wellness, instead of treating them as separate issues.
“People don’t just ask about physical health issues. They also ask things like, ‘I wear a hijab and I want to go for a run, but I feel really awkward,’” said Russell. “It sounds like a very normal question, but it’s actually quite a loaded question, because it’s affecting their mental health on a day-to-day basis.”
One of the company’s goals is to make the app feel accessible, so people feel more comfortable seeking support. “We’ve literally have had sweets delivered to our office when a user has a baby,” Russell said. “These are the personal touches that I think Maya has delivered in terms of dealing with both physical as well as mental health conditions combined together.”
The company is currently working with different monetization models. One is business-to-business sales, positioning Maya as a software-as-a-service platform that employers can offer to workers as a benefit. Garment manufacturing is one of Bangladesh’s biggest export sectors, and many workers are young women, fitting Maya’s typical user profile. The startup has worked with Marks and Spencer, Primark and the Bangladesh Garments Manufacturer and Exporters Association (BGMEA).
Another B2B route is partnering with insurance providers who offer Maya as a benefit. On the direct-to-consumer side, Maya recently launched premium services, including in-app video consultations and prescription delivery. Demand for consultations increased sharply during the COVID-19 pandemic, and it now handles about 300,000 video visits a month. Russell expects many users to continue using telehealth services even after the pandemic subsides.
“They’ve really seen the advantage of just having a doctor right in front of you,” she said. “For people with chronic conditions, it’s easier because they don’t have to go somewhere every week, and the fact they have monitoring and their history gathered is helpful for regular users, too.”