As India hunts for COVID-19 resources such as oxygen cylinders and hospital beds, technologists leverage conversational AI — such as Introbot AI and CovidIndResBot — to create an online space for real-time responses with a touch of personability
(Subscribe to our Today’s Cache newsletter for a quick snapshot of top 5 tech stories. Click here to subscribe for free.)
Utkarsh Roy is on the road to recovery from COVID-19 in his home in Chhattisgarh. He feels exhausted with the constant news of new cases in his immediate circle. But the engineer powers through, as he and his Introbot AI co-founder, Divyaansh Anuj, are working round the clock updating and streamlining this WhatsApp-based conversational Artificial Intelligence that has been repurposed as a COVID Helpline bot to help people find verified critical medical resources such as oxygen cylinders and hospital beds in real-time across India.
Meanwhile Twitter bots, typically seen as a nuisance, now include useful ones. COVIDResourceIndia (@CovidIndResBot) on Twitter was built by Rahul Raina and Naman Gupta who are students of School of Accelerated Learning, Hyderabad and Mumbai.
They were guided by SOAL CEO Raj Desai as well as volunteer-driven tech community, The Product Folks, and tourism channel, Travel XP. The COVIDResourceIndia bot too, has been helping people find hospital beds, ambulances, oxygen suppliers and medicines, and responses are in real time.
Chatbots have long been regarded with a side eye for a number of reasons; you cannot hold a chatbot accountable if something goes wrong. But in the last year, these AI-powered entities have proven useful for their scalability particularly during the pandemic.
- MyGov Corona Helpdesk: This WhatsApp-powered platform created by the National eGovernance Division is helping people find their nearest vaccination centre and book appointment slots. An earlier version was launched last year to help eradicate fake news about the virus.
- Yellow Messenger: This Omnichannel chatbot uses any text platform for COVID-19 use cases, offering help with vaccine registrations, oxygen/bed availability, collecting patient data, mental health assistance and insurance queries.
- International Fact-Checking Network: In June 2020, a Hindi version of the IFCN conversational AI was launched to bust myths around the Coronavirus. This bot sources verified content from major news organisations and is updated daily with 250-plus fact checks.
With everyone needing a database of verified COVID leads, the problems are universal, but the solution offered by chatbots has been unique. One of the first chatbots that emerged during Coronavirus was a WhatsApp-enabled chatbot commissioned by the World Health Organization (WHO), in helping people understand the symptoms for the various virus strains. This paved the way for stronger conversations and more development.
Says Utkarsh, “The problem around chatbots was not with the interface itself but with the limited application and intelligence behind these, as chatbots did not do much aside from being front-end of the inovation of information given. We have built a central intelligence that could connect different communities, businesses and individuals; it wasn’t designed to be just a chatbot but its interface is one.”
But with existing central intelligence connecting virtual communities, Utkarsh and Divyaansh were able to develop and deploy a response system in a short amount of time, as the second wave hit India. Utkarsh says they already had a plug-in with sizable communities across the country (VC portfolios, startup communities, ISB, IITs and IIMs) and were able to reach as many people as possible already. This COVID initiative of Introbot AI, states Utkarsh, is a not-for-profit and community-led. He adds, “Every rupee goes towards maintaining our servers round the clock.”
COVIDResourceIndia being a Twitter-based entity is no small solution, given how social media has proven to be a helpful amplifier of verified information.
Raj, who calls Gen Z the “problem solvers”, says, “[The Coronavirus] has led to a simultaneous outbreak of misinformation. Such unverified information prevents people from getting timely help and support. When it comes to the recovery of a patient, even an hour of early admission can make all the difference… Rahul and Naman [the developers] have taken the fight where it matters and this will go a long way in saving several lives across India.”
Utkarsh says time and scale are major factors in developing and perfecting Introbot AI for public use, and he says it continues to be so. “Divyaansh and I are serial founders and are used to forecasting what is around the corner and how we can help there,” says Utkarsh, who has experience building multiple startups in the past and large cloud-based systems. Divyaansh, he says, has experience in the creator space of building a community of 50,000-plus professionals.
“Everyone in India is thinking the same thing right now: how can we build a database of verified COVID leads. We went from serving 200 COVID-affected families on day one to half a million in seven days, so we were inherently focussed on scale to begin with and was not an afterthought,” he insists. “We also partnered with COVID Citizens, a community initiative by the startup ecosystem in India and an alliance of 15 such partner organisations such as NowFloats, WakeFit, United for Social Impact, Vruksha Ecosystem and others across the country. FireSide app, started by the founders of Chingari [a short video app] is also on board, as well as 350-plus volunteres on the ground.”
Working in a crisis situation is a race against time and anything less would be a liability. The Introbot AI team — which has seen 100 to 150 users engage with the platform by the minute — recognises that if someone were to request 10 oxygen cylinders in Kolkata, they would run out very quickly. So the database has to be constantly updated.
Utkarsh Roy and Divyaansh Anuj, co-founders of Introbot AI
“The dynamic in drawing these updates from 300-plus cities on the ground has been a massive challenge. We are also doing load balancing on our back-end. For every medical supplier there are 200-plus people looking for that supply, so we have to make sure the supplier gets a balanced number of requests a day. Fraudulent and unavailable suppliers are constantly being removed from the system. We also have a unique crowdsourcing system where anyone can add supplier information. Through this, we have one of the largest medical supply databases (with 25,000 suppliers) in the country.”
Naman concurs, explaining that working on the COVIDResourceIndia offered him insight into working on “a cross-functional, fast-paced ecosystem to decrease the turnaround time in product development.”
Introbot AI metrics
- Has observed 25,000-plus people finding medical supply information successfully
- 100% response to all queries in under two seconds
- 300-plus cities served across India
- A medical supply database with around 25,000 suppliers
- The platform was built in seven days
Utkarsh and Divyaansh chime in saying that during the pandemic Indian tech and business circles have seen unprecedented collaboration on every front. “Now is the time to put aside personal motives and come together. When we see people on social media or in private messages saying how Introbot AI has been helpful in saving a life, we are reminded of the power of technology for good. In the coming future, we may see Introbot AI connecting people with one another in varied ways; for example, post-COVID, people may seek counselling services in a telemedicine space.”
Rahul agrees, “We truly believe that tech can be a force that can be used for good, especially when India’s strength is compromised in both numbers and resources. We are doing our bit to relieve this pressure with the development of COVID resourceful chatbot and maintaining the authenticity of its database.”
Some who have already used @CovidIndResBot and Introbot AI would have noticed a more personal and conversational tone. Language has been an integral element in developing the interface for bots such as @CovidIndResBot and Introbot AI. Utkarsh says, “Creating impact in quantity is not the only goal but also through emotion. Each word we place out there can be helpful. Our previous central intelligence hub was too conversational actually, so we had to tone it down and adapt it for all tiers of society for this rollout.”