Published in


Quora AMA session by Aakrit Vaish

“I sleep pretty well knowing that we have a team that is always a step ahead of me”

answered Aakrit when asked what keeps him up at night while building Haptik, part of one of the most candid AMA sessions you will ever see. Hosted on Quora, there are close to 30 such answers on Chat Bots, AI, technology, entrepreneurship, and of course, Haptik.

Here are some of our favorite picks.

What are some of the biggest turning points in your journey building Haptik?

1. March 31, 2014: We had a really unsuccessful beta phase of Haptik in December 2013. Based on all available logic, we were going to drop the idea completely. We even came up with 4 alternate ideas. However, the rest of the team wanted to give one final real shot to the product. We launched the Haptik app on March 31, 2014 and the first 24 hours saw 5x more traffic than we expected. We knew we were on to something.

2. August 2015: We had been pitching to investors for 3 months for a Series B round. We pretty much had a verbal go ahead from one of the largest investors at one point, but it all fell apart. We were 2 months away from running out of money when I landed a chance meeting with Satyan Gajwani. He showed great enthusiasm about how Haptik could fit into the larger strategic vision of Times Internet. One thing led to another, and the opportunity seemed larger than what we had ever thought of with any other VC. We closed our Series B with Times Internet in April 2016, giving a 2x return to our earlier investors.

3. February 2017: We had been doing a lot of work over the last few months on the technology platform. We always knew the enterprise use case for chatbots would be huge, but for some reason did not dive into that business. On Feb 13, we got an inbound email from an industry acquaintance that Coca-Cola wanted to meet with us for a chatbot idea. We met them the next day at Starbucks Mumbai Airport, and they gave a brief for a chatbot they wanted to make for their Facebook page. We had never built a Facebook bot, but we gave them enough confidence to bet on us. Since then we have signed on 12 more clients of similar size and our enterprise business is the fastest growing business line in the company.

4. Today: We just launched an integration with the Times of India Android app, where Haptik is embedded inside the app. This makes Haptik instantly accessible to millions of users without the need to download it separately. This is a technological breakthrough where more than 20 services and 6 PGs/wallets are integrated inside a news app for the first time globally, all in under 1 MB. We expect our monthly active users to double every month from here on for the next 6 months.

Bots save money for companies, but do humans really want to interact with chat bots? Why?

This is a great question. The jury on chatbots is still out there. As the biggest evangelist of the chatbot space, I will be the first one to honestly say that it is not 100% proven yet that humans do really want to interact with chatbots.

But I can tell you there is a great degree of belief that eventually people will latch on to chatbots. Why?
It is a very lightweight interface to learn and understand. Chatbots don’t have menu buttons or too many screens to navigate through. It is a simple send-receive message interface that any person can pick up and understand within a few minutes.

It is the most natural interface for the smartphone. Mobile apps are simply desktop websites re-fitted for the smaller screen. But messaging or chat is ground up a smartphone first use case. 75% of our daily usage on our phones is inside a messaging interface. As this becomes the default, services will also have to move to messaging. Chatbots are not necessarily machines masquerading as individuals, but rather an interface.

With a chatbot, it’s a lot easier to personalize the experience than a traditional experience. The bot can learn about your preferences and accordingly change the direction of the conversation. Also, each chatbot can have a unique character to it making the experience of getting something fun and unique to each brand.

We look at this as a paradigm shift large as when the world moved from a command line interface to a GUI. It will take time for people to get used to doing tasks with chatbots, but it will happen.

What is one of the most important things you’ve learned from your entrepreneurial journey thus far?

The only thing that matters in entrepreneurship is perseverance. I live by this quote from the movie The Founder (story of the McDonald’s founder): Nothing in this world can take the place of good old persistence. Talent won’t. Nothing’s more common than unsuccessful men with talent. Genius won’t. Unrecognized genius is practically a clich. Education won’t. Why the world is full of educated fools. Persistence and determination alone are all powerful.

How does Haptik or any chat bot company monetize today?

We have multiple revenue generating business lines:
1. Commission on e-commerce transactions such as flight bookings, product sales, cab bookings, etc.
2. Native chat bot driven brand advertising.
3. Software-as-a-service for custom bot solutions for large enterprises.

What are some of the most popular requests you get in Haptik? How do you see that changing?

Currently, our most popular request is for Reminders. This validates our theory that bots are great for simple tasks. It’s a LOT easier to simply send a message saying “Remind me to call Tanya”, than opening an app, filling up details, setting a time, etc. We just launched Jokes on Haptik last week which is seeing some great traction. We expect simple day to day stuff to keep growing in volume. But even our transaction categories particularly Flight booking continues to be very strong.

Oh, and we do more than 100 web check-ins per day!

What machine learning frameworks and tools do Haptik use?

Haptik uses a combination of Natural Language Processing and Machine Learning to build most of it’s algorithms. Apart from our custom algorithms, we use functionalities from python NLTK, Gensim word2vec, Stanford NLP for Natural Language processing and Scikit-learn, Tensorflow, for learning algorithms.

Don’t you think Haptik is useless in the world of Siri and Google now? Exempting the discounts they give for using them.

Good question. Intuitively, even I would have thought so. But I don’t know anybody who has gotten anything really done by Siri or Google Now. Like, have you ever managed to get an end to end task completed via any of them? They work really well for search, but not for task completion.

How do chat bots handle multilingual requests?

To build a great chatbot, you need to write a great script. That cannot be done by simply translating English into Hindi. That requires thinking of it ground up in the native language. It’s as much a product design challenge as it is a technology challenge.

There are too many chatbot companies of different kinds. How do you differentiate yourself?

We are one of the only companies in the world that owns the entire product technology stack of building a chatbot. Additionally, we are also one of the few companies that do business on both the consumer and enterprise side.

Where do you want to see Haptik after 10 years?

Hopefully, in 10 years,we are all still alive! If we are, then I hope that by then AI is not just a trend, but a day to day reality and Haptik is one of the foremost companies globally in the space.

As a leader, what’s your way of grooming employees into leaders?

Give them clear, specific quantifiable goals and then leave them alone. Up to them on how they want to achieve them.

Click here to see the entire session.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store



Global leaders in Conversational AI, powering Intelligent Virtual Assistants (IVA) that transform Customer Experience