Betterhalf: The millennials of India are turning to algorithms for love
Spearheaded by Pawan Gupta, an MIT Sloan School of Management alum, Betterhalf uses AI to determine emotional, intellectual and social compatibility.
Pawan Gupta wasn't having any luck finding a match.
The 31-year-old MBA graduate wanted to get married and scoured dating apps and Indian matrimonial websites. Yet after four years of looking and dating, he was no closer to walking around the sacred fire typical of the Hindu marriage ceremony.
So the MIT Sloan School of Management alum decided to build an AI-powered matchmaking app, one that would determine emotional, intellectual and social compatibility. Gupta teamed up with engineer Rahul Namdev to create Betterhalf, employing a wide swath of data to figure out who could be successfully matched for marriage.
The service is starting to have success. Manoj Kumar Singh, 32, turned to Betterhalf after years of failures with traditional matchmaking and dating websites. He was paired up with an HSBC executive he found kind and caring; they were married within months. "Over four years, I thought I'd tried everything," Singh recounted telling friends when he first heard of Betterhalf. "After so many frustrating attempts, an AI-based app helped find me a wife."
AI systems are flourishing in dating apps the world over, using deep-learning to analyse Facebook posts or users' tweets to gauge how compatible they would be for each other. Nowhere is this more significant than in India, where marriage remains a goal for many in the world's largest youthful population of more than 400 million millennials. Local entrepreneurs, global dating apps and Indian matrimonial websites are racing to bring artificial intelligence to modernize match-making.
There is, however, a lingering question. "Is India culturally ready for AI matchmaking apps?" said DD Mishra, a research director and analyst based in India for Gartner Inc.
Gupta's experiences with Indian matrimonial websites and global websites suggest young urban Indians are ready for something new. He had become increasingly distressed by the process. One match made over a site went awry when she told him she was interested only in a spouse who could get an overseas posting. On another first date, he arrived wearing a new shirt to meet a woman matched by a matrimonial website who left him stranded. And it took him months to realize that someone he was chatting with on a dating app wasn't interested yet kept him dangling.
"I felt rejected again and again," Gupta said. "There's a woman out there for me, but none of the dating apps or matrimonial websites was helping me find her."
So he turned to algorithms and started Betterhalf.
The firm, funded by 15 angel investors including half a dozen MIT alumni, puts users through several levels of verification comprising social networks, phone number, personal and work email and government ID. New registrants answer a breadth of questions including on moral values, emotions and personality traits. That information is then melded with data from married individuals, open-source data and compatibility research. After matching people, post-interaction user polls help weed out the creeps.
There are thousands of dating apps around the world, Gupta said during a conversation at the Betterhalf's headquarters, a home converted into quirky office space. "There's a ready market of tens of millions of educated urban Indians seriously looking for partners so why is nobody getting it right?"
Betterhalf is targeted solely to serious spouse seekers. The startup has 15 employees -- 13 are single, use the app and half-mockingly refer to their office as a marriage lab. They are in the app's core demographic, ages 25 to 40. Product head Rohan Bhatore, 27, has been avidly searching for seven months. The engineer educated at the Indian Institute of Technology has been matched with a dozen prospects but hasn't yet advanced to a meeting. "Searching for a marriage partner is frustrating and soul-numbing," he said.
Algorithms can end up knowing a person better than friends, family or even themselves, and that's revolutionizing matchmaking, said Michal Kosinski, a computational psychologist and assistant professor at Stanford University's Graduate School of Business who recently signed on as an adviser to Betterhalf. "Algorithms can learn from experiences of billions of others, while a typical person can only learn from their own experience and the experience of a relatively small number of friends," he said.
This is a much different matrimonial approach from the ones Indians have known for centuries: Matchmakers or families find prospective spouses within their circle and then vet the duo's horoscopes by astrologers to ensure that the stars are aligned for the match.
There's plenty of competition in the new AI space. Delhi-based entrepreneur Kumar Akshay's AI-powered startup Truematch, which uses LinkedIn to authenticate users, is trying to dispel the idea that matchmaking apps only lead to casual relationships. "We wanted our app to be as trustworthy as the family matchmaker yet as cool as a dating app," he said.
Even some of India's traditional matrimonial giants are jumping on the bandwagon. Matrimony.com, which has 3.7 million active users, is developing an AI-powered chatbot that it plans to deploy later this year. And global apps like OkCupid and Bumble, backed by Bollywood-Hollywood star Priyanka Chopra, are going beyond matching with obvious criteria like language or vocation in India, and getting in tune with cultural realities.
"AI works best in matchmaking when you have large volume and variation of data, which India offers," said Meenakshi Variankaval, who heads data sciences at Matrimony.com.
AI capabilities also could help detect matrimonial fraud, a serious challenge in India. In one recent case, a serial con artist was arrested in Hyderabad after he created dozens of fake profiles. He married one woman, was living with another and had defrauded hundreds more of millions of rupees.
"AI has the processing and cognitive capabilities to help matchmaking apps carry trust indicators on particular profiles," said Mishra, the Gartner analyst. "This can generate huge value as it addresses the biggest risk in matchmaking."
Betterhalf has other hurdles. In a diverse country, many users come up with extraordinarily specific, inflexible partner preferences. "How do you design a product for very busy professionals by giving them what they want and eventually leading them to relationship happiness?" Gupta said, adding the app was at the "2 percent level" of the technology it eventually wants to build.
Gupta could be his own lab rodent. For a few months he has been dating a management consultant he was matched with on Betterhalf. They've found they are compatible in areas like managing finances and sharing household chores. Both sets of parents approve and Gupta said his search may soon be over. "An invite to my own AI-engineered wedding could be around the corner," he said.