She was looking for a specific dress - the kind that has a name in the vocabulary of any woman who wears it. Something like "a midi sundress with a flutter sleeve for a garden party." The search bar returned results that had nothing to do with what she had in mind. The product descriptions lived in a different language than the search query. Retailers called things what they called them; shoppers called the same things something else entirely.
Gupta's first instinct was to blame herself. Perhaps this was a translation problem - an immigrant unfamiliar with American retail vocabulary. So she did something most people would not do: she went and asked 1,000 other women. All of them had the same experience. The disconnect was not cultural. It was structural.
That structural flaw became the foundation of Lily AI. Founded in 2015 alongside co-founder Sowmiya Chocka Narayanan, the company set out to build what Gupta describes as the emotional intelligence layer of retail - technology that understands not just what a retailer calls a product, but what a shopper feels about it, how they'd describe it, and how they'd search for it.
Before building her own company, Gupta had spent years at the intersection of brand and behavior. Her career began at Saatchi & Saatchi, where she worked on P&G accounts and developed what she later called a fascination with "the emotional strength of the connection people had to their preferred products." She moved from advertising to social impact, joining Eko - a Gates Foundation-funded branchless banking platform for the unbanked - and then UNICEF Ventures' Innovation Fund, where she invested in life-saving applications and technologies.
The career arc looks eclectic on paper. In practice, it produced a founder with an unusual combination: deep consumer empathy, comfort with large-scale systems, and an eye for where technology meets human need. That combination is precisely what Lily AI required.