Presenter: Ankush Sabharwal

Moderator: Jibu Elias 

This week's session was led by Ankush Sabharwal, cofounder of CoRover. Many of you may not know right off the bat what CoRover is but you would have definitely come across AskDisha - the conversational AI platform for the Indian Railways or IRCTC. CoRover is the company that has built this cognitive AI technology platform , and it offers managed Chatbot as a Service, with Self-Onboarding support, to help enterprises generate revenue, save cost, improve customer experience and operational efficiency. CoRover has garnered more than 200 million unique active users, with an excess 4 million daily active users (DAU) and 500,000 concurrent users. CoRover has the highest traction in India - nearly 2.5x as compared to the closest competitor.

CoRover’s main competitive advantages are:

• Versatility: Multi-format - video, voice, text Omni-channel - Web, App, WhatsApp, IVR, SMS, Social Media Bots) and Multilingual, since it works with 12 vernacular & 100 foreign languages
• Tech Edge: Robust, Secure, Scalable Architecture (>500K concurrent users supported) Designed grounds up to be deployed on public and private clouds and on premise. An award-winning platform
• Turnaround Time: 10x quicker to deploy, thanks to ease of training with our proprietary patent pending CBML (ChatBot Markup Language) technology. Train the bot in English and make it available in any language with industry leading (>95%) accuracy.

CoRover works with several large enterprises like Microsoft, Birlasoft, Google, Accenture, AGC, ATC/ITC, IGL, Automation Anywhere, UI Path, IRCTC, Flipkart, Orange, KSRTC, NPCI, AIM and NITI Aayog. 

How Will Chatbots Ensure Business Process Continuity? 

In the current scenario, human interaction has significantly dropped although the consumer queries have not. In some cases, the extent of consumer queries have actually increased such as governance and healthcare. In any sector, problems need to be heard and consequently solved. At times like this, it can get very hard to asnwer all queries especially when multiple communication modalities exist - we have phone calls, emails, chatbots, IVR and even voice IVR. It would be hugely beneficial to have all queries answered by one single channel of communication. We believe that voice can do this very efficiently. 

The next set of challenges to address are immediacy and authencity of information, and in multiple languages. When it comes to a written query, the natural response would be English but this applies to a small percentage of the Indian population. We have several vernacular language speakers, and to build a bot that can respond to these queries would help greatly. 

With chatbots, we have realised that nearly 70% of queries are standard (based on sector) and repetitive. By using voice and training the engine using NLP, and adding AI to understand the queries and process responses, the experience of a machine communicating to a man and successfully solving a problem can be established. Now is the most ideal time to integrate and migrate to voice-driven bots, since call centers are not running to full capacity and probably will not for a considerable amout of time. 

Understanding Use Cases To Build Bots With Value 

Mere application of a chatbot to a business model isn't going to guarantee success. The business owner needs to analyse the application of bots and see how it deliver maximum value. For instance, one of the more popular use cases we have is the integration of live refund status through a customer's PNR number for IRCTC. This is in addition to seeing train schedules, ticket status, refund status and more. This new feature is directly addressed through the app, in either Hindi or English. Another use case is with Indraprastha Gas Agency in New Delhi, where we have enabled customers to check their outstanding bill amount through AskMythri (a bot we provided) and to raise a ticket through the app if there is a gas leak, which will alert the service executives and have them send someone over to the house within two hours to fix the problem. One of our newest value additions is PAi, an AI chatbot for enhancing awareness on digital payments especially RuPay, FASTag, AePS and UPI in realtime. PAi will also be available in several regional languages in India which will increase the adoption of digital payments among the Indian citizens.; users will be able to get verified automated responses to their queries on NPCI products with the help of PAi and will also be accessible to global RuPay cardholders.

One of most successful use cases was AskDoc.ai, which we developed as a rapid response to COVID19. With Askdoc.ai, data provided by WHO, MoHFW, GoI and ICMR is gathered, and Artificial Intelligence and NLP techniques are utilized to dispense it in an easily digestible manner. In addition, they have brought on board a doctor – Dr. Pratik Patil, a physician specializing in infectious diseases at Bangalore’s Fortis Hospital – to be able to speak to people, adding a much-needed personalized element to providing information. The responses are automated, to commonly asked questions about COVID19 with specifics derived from firsthand data from the Ministry of Health & Family Welfare. This feature was developed keeping in mind how Indians are comfortable procuring treatment from a known doctor, and seek expert help based on familiarity with a physician. 

The logic is fairly straightforward - to build better bots, you need more data.