Some time in early 2016, Siddhartha Pahwa thought it would be “disruptive” if he could make a software that will allow users to book a cab on WhatsApp.
That would give Pahwa’s Meru Cabs a technological edge over other cab aggregators such as Uber and Ola, but to his disappointment WhatsApp did not allow automated messaging.
But on April 12, when Facebook opened up its Messenger for developers to build similar applications, Pahwa, who is the CEO of Meru, built an application on Messenger, which allowed India’s 150 million Facebook users to book a Meru cab through a simple chat. It just needed a one-time verification of the customer’s phone number.
On the other side of the chat is an application called a bot, which is intelligent enough to understand human language, send back the query to the company’s central server, find a cab and send the booking information back to the customer.
In just a couple of months Pahwa has seen more than 300 people using the Meru chatbot, with 35 daily bookings. “Chatting comes naturally to people,” says Pahwa. Meru’s Messenger bot chats like a human. When you apply machine learning, he says, the bot starts to distinguish more.
Bots can not only hail a cab, but give news updates, book appointments, manage work calendars and even help in shopping. Recent news articles show 24% of all messages on microblogging platform Twitter are made by bots. The Guardian reported that half of the clicks on internet aren’t made by humans any more, and that 22 out of 30 most prolific editors on Wikipedia are actually bots!
Global tech giants have put their weight behind bots. Facebook’s platform is called “M”. Google’s Now, Apple’s Siri, Microsoft’s Cortana and Amazon’s Alexa are voice assistants that combine spoken language with artificial intelligence (AI) to deliver similar services. Enterprise software company Slack’s Slackbot acts as an office assistant.
Flock, the Indian counterpart of Slack, also uses a bot that notifies the user when a work task gets completed. For example, it can create a to-do list, make announcements and set reminders — or almost everything that an office assistant can do except getting you a cup of coffee. To be sure, though, it will be able to order it. “With AI, bots will be able to do the full-fledged work of an assistant using natural language processing,” says Bhavin Turakhia, founder and CEO of Flock.
Many companies in India are making bots. Healthcare app iCliniq has a Slackbot and a Telegram bot. Dhruv Suyamprakasam, founder of iCliniq, says he has trained them to recognise 4,500 words and a lot of data. “Training a bot is not easy. A query with the word acne means it is for a dermatologist, varicose veins should take the query to a vascular surgeon,” he says.
Though millions of apps have come up in the past few years, an average smartphone user may keep 10-20. Facebook, Messenger, Twitter and a couple of shopping apps take up most of the space. “There is app fatigue. People don’t download them… the top ones are the messaging apps,” said Beerud Sheth, founder and CEO of messaging platform Gupshup.
Not everyone believes bots are the future, though. Vijay Shekhar Sharma, founder and CEO of Paytm, says what Facebook builds cannot be trusted. “They will shut it down anytime... they want to control the internet,” he says.
Others say the use of bots is limited to messaging. In other cases, websites and apps offer far more control than bots. “If you want to do that on a bot, it will become a force fit,” says Sheth.
Apart from Meru not many e-commerce companies have warmed up to the concept. Flipkart, Snapdeal, Paytm and ShopClues don’t have bots — yet.
But one should not forget that bots learn from real people. Microsoft’s Twitter bot Tay backfired as it modified its language according to what users were tweeting and became everything from sexist to racist. While this is a testament to the possibilities opened up by bots, it reasserts the fact that they need to be created carefully. There is no substitute for good upbringing!