‘Revenge dress for a party in Sicily': This platform is using AI to make online shopping hyper-personal

CNN

4 days ago

As anyone who's scoured the internet for a bridesmaid dress knows, online shopping can be a pain. Among almost unlimited options, it can be a difficult task to find just the right style, color, size and price point.
A tech startup called Daydream is now looking to fix that by letting users search for a product online in the same way they'd describe it to a friend.
A user could say they're looking for a 'revenge dress to wear to a party in Sicily in July,' for example, or 'a summer bag to carry to work and to cocktails after.'
Daydream, which has staff in the New York and San Francisco areas, is just the latest tech company using artificial intelligence to try to make online shopping simpler and more personalized. The demand is already there — a survey of 5,000 American consumers published by Adobe Analytics showed that 39% of respondents had used a generative AI tool for online shopping last year and that 53% planned to do so this year.
It's competing with tech giants that have already launched AI tools for online shopping. Meta is using AI to make it easier for sellers to list items for sale on its apps, and to show users ads for products they're more likely to buy. OpenAI launched an AI agent that can shop for users across the web, and Amazon is testing a similar feature. And Google has rolled out a range of AI shopping tools, including automated price tracking, a 'circle to search' feature that lets users search for a product in a photo or on social media, and virtual try-on for clothes.
But Daydream has a deeper understanding of the fashion and retail industries than those bigger players, CEO Julie Bornstein told CNN. Bornstein helped build Nordstrom's website as its vice president of e-commerce in the early 2000s and worked in the C-suite for Sephora and Stitch Fix. In 2018, she co-founded her first AI-powered shopping startup The Yes, which sold to Pinterest in 2022.
'They don't have the people, the mindset, the passion to do what needs to be done to make a category like fashion work for (AI) recommendations,' Bornstein said. 'Because I've been in this space my whole career, what I know is that having the catalogue that has everything and being able to show the right person the right stuff is what makes shopping easier.'
Already, Daydream has raised $50 million in its first round of funding from investors including Google Ventures and model and Kode With Klossy founder Karlie Kloss.
The free platform operates sort of like a digital personal stylist. Users can type in what they're looking for in natural language — no Boolean search terms required, thanks to its AI text recognition technology — or upload an inspiration photo. Then, Daydream will surface recommendations from more than 8,000 brand partners, ranging from Uniqlo to Gucci. Users can then continue chatting, just like they would with a chatbot, to refine the search; for instance, by asking for more casual or less expensive options. As users spend more time on the platform, it will start to tailor recommendations based on what they've searched for, clicked on and saved.
When they're ready to buy, shoppers are directed to the brand's website to complete their purchase, and Daydream will take a cut of the sale. Unlike many of the other big players in e-commerce, Bornstein is eschewing ads-based rankings — she wants products to show up on recommendation pages because they're a likely fit for the customer, not because brands have paid for them to be there.
'As soon as Amazon started doing paid sponsorships, I'm like, 'How can I find what the real good product is?'' she said. 'We want this to be a thing where we get paid when we show the customer the right thing.'
On a recent CNN test of Daydream, a search for 'white, fitted button-up shirt for the office with no pockets' led to a $145 cotton long-sleeve from Theory that fit the bill. But the recommendations aren't always perfect — a search for 'mother of the bride dress for a summer wedding in California' returned, among more formal styles, several slinky slip dresses, including in white, that seemed more suited to a bachelorette party.
Bornstein said the company continues to refine its AI models and collect user feedback.
'We want data on what people are doing so we can focus and learn where we do well and where we don't,' she said.
Part of that work, she added, is training the AI model to understand what it means when users say, for example, they're looking for a dress for a trip to Greece in August (it's going to be hot) or that it's for a black-tie wedding (it should be formal).
Daydream launched its web version to the public last month, although it remains in beta testing, and plans to release an app this fall.
In the future, Bornstein said she expects people to use AI not just for shopping but for a range of fashion needs, such as pairing items they're shopping for with existing pieces in their closet.
'This was one of my earliest ideas, but I didn't know the term (generative AI) and I didn't know a large language model was going to be the unlock,' she said.