Become a Vogue Business Member to receive unlimited access to Member-only reporting and insights, our Beauty and TikTok Trend Trackers, Member-only newsletters, and exclusive event invitations.
The Daydream team has a refrain that’s guided the AI shopping platform’s user experience since it began building: Say more.
Looking for a red dress? Say more. Is it for a wedding? A work conference? Where will you wear it, and what time of year? Do you like a low cut or a high neck? Cocktail length or gown? Be as specific as possible, and as you circle in on what you’re really looking for, Daydream’s artificial-intelligence-powered, chat-based shopping agent will generate relevant results.
This say-more style of shopping is what Daydream believes will change the way we find what we want to buy. Julie Bornstein, Daydream CEO and founder, has been in the business of search for 25 years, including those spent at Nordstrom in its early e-commerce days, then Stitch Fix, Sephora, and her first startup, The Yes, which was sold to Pinterest in 2022 before being subsequently shut down by the social platform. The Yes was a sort of predecessor to Daydream, Bornstein concedes, but large language modeling and AI technology have come so far since she launched that company in 2020 that the user experiences are completely different, she says. Last year, Bornstein raised $50 million to build Daydream, coled by investors Forerunner Ventures and Index Ventures, with a founding team including chief brands officer Lisa Yamner Green, chief product officer Dan Cary, and chief strategy officer Richard Kim.
Today, Daydream is going live with a beta version at Daydream.ing that Bornstein says is ready to start providing product recommendations using its generative AI shopping tool. An app will come later this summer. Users answer a set of questions to create what’s called their style passport, such as what brands they like and what sizes they wear, among other shopping proclivities.
Then they can start chatting. Daydream’s AI interface will be familiar to anyone who’s experimented with ChatGPT, but its overall feel is more e-commerce retailer meets Tumblr feed. As you tell the AI what you’re looking for, Daydream will pull matching results from its 8,500 brands (some listed directly, others through retailers) and more than two million products, across womenswear, menswear, and accessories, with launch partners including Net-a-Porter, Alo Yoga, LoveShackFancy, Markarian, Khaite, Emilia Wickstead, Mytheresa, Cult Mia, Mejuri, Uniqlo, and Dôen.
Amassing more products is a priority: These brands are all searchable on Google or pullable by ChatGPT. The product selection is table stakes: Daydream believes its competitive advantage will lie in not only having the largest catalog but also the best understanding of shopper intent and its ability to get the right products in front of the right people.
“Search in fashion has never been solved,” says Daydream chief technology officer Maria Belousova. “It’s about discovery and personalization”—the ability to read between the lines of search prompts.
That’s where the AI comes in. From initial results, you can upvote or downvote items as well as further narrow the search, specifying style, color, material, or situation or selecting an item from the feed to tweak it. Photos can also be uploaded to the chat to prompt the search for a similar—or the same—item. If it finds you the perfect pair of work trousers but you want them in wool, you’ll write that in the chat to generate more tailored results. There’s no universal cart or in-platform checkout, though Belousova says agentic AI could remove some of the hassle of manual checkouts further down the road. When you want to buy something, Daydream links you out to the brand or retailers selling it and charges a commission fee of 20%.
Daydream’s release comes as conversational AI and AI-enabled shopping have reached new levels of mainstream adoption. OpenAI’s ChatGPT has become part of many people’s everyday lives, and in April the company introduced the ability to purchase items directly in the chat interface. Google has invested in bringing AI to the forefront of its search results and in May announced the addition of AI Mode to its shopping tab, which offers a generative AI chat function to help users find the products they’re looking for. Alta, an AI discovery platform founded by Jenny Wang, recently raised $11 million to build out its shopping capabilities. Now Daydream has entered the chat.
“We want to make it easier to find the things you love in the world of fashion,” says Bornstein. “Google has been at it for a long time, but there’s a lack of understanding in the category. ChatGPT is training people to ask for things differently, but we have the fashion expertise.”
Can AI fix online fashion discovery?
These developments all point to a future where shopping online will no longer be limited to standard e-commerce filters and searches that can only generate results around typical product tags, like color and style. Bornstein says Daydream’s process of onboarding items with robust descriptions starts with how a brand has already tagged it. Then Daydream runs every item through its AI model, which attaches more descriptions around objective and subjective features: Would a linen dress be good for travel? What season is this specific floral dress appropriate for? Brand mapping—the web of brands that Daydream will serve to a user, recommended based on their responses to style questions and searches—will play a key role in both brand discovery and personalized results.
“We’ve spent the last year working on this robust chat-to-search capability,” says Bornstein. “The technology has continued to evolve, and the way you build a consumer application on top of AI has evolved. We found ways to improve our accuracy and our speed, so that’s been a big area of focus for us, as well as a rich understanding of the products.”
The goal is that customers—with enough prompting—can bend search queries to the exact item they’re looking for, making them more likely to purchase. While brands will eventually be able to add editorial, among other features, to their brand pages, the search results will never be pay to play, Bornstein says. Daydream doesn’t run ads; it only makes money from each sale made via the platform.
If successful, Daydream’s AI agent could help solve a problem facing brands and fashion lovers alike: Platforms like Google and Meta’s business models are driven by advertising. Brands need to pay the highest bid to get in front of audiences; customers, in turn, have lost trust in the algorithm for shopping inspiration.
Bornstein believes the problem is that the right algorithm has yet to be applied to fashion search. “When I saw ChatGPT and understood what an LLM [large language model] could do, I knew it was what we needed to make fashion search and discovery possible online,” she says. “That personalized stylist conversation has not been possible to date, and suddenly it is.”
Brands are also looking for alternative ways to get in front of customers as acquisition costs climb. Holly Soroca, president of womenswear brand Dôen, sees AI-chat-based shopping as an opportunity to access customers who are looking for something specific, with high intentionality. “Searching for a floral dress and having to scroll through endless options without curation is really difficult in the multi-brand e-commerce shopping experience,” says Soroca. “Having an iterative dialogue of what the customer is looking for, beyond just a dress—it connects with what they’re really emotionally looking for.” Soroca says that Daydream as a discovery engine will complement the way the brand already shows up on platforms like Instagram and in multi-brand retail.
As Daydream launches in beta, Bornstein and her growing team are already testing a second algorithm against the first to improve the personalization and accuracy of results. The early test runs of this type of shopping experience are critical: If users don’t feel the results are matching up to the promise of better personalization, they’re likely going to be quick to opt out.
Early stumbles, like linking to a product on a brand’s website that’s no longer available, will need to be ironed out. The other challenge is getting them to try Daydream in the first place: Customer acquisition is as hard for a new shopping platform as it is for a brand. Bornstein says early marketing plans include working with stylists to create collections that users can then look to for inspiration, but the early days will depend mostly on word of mouth and PR.
Bornstein says she hopes that customers—many of whom are frustrated by current shopping experiences—will join for the ride.
“People have very high expectations,” she says. “When you’re a first-time user, it’s such a delicate moment. We want you to have a great experience.” It won’t be easy. “Fashion is so complex. But it’s been my lifelong obsession, and we’re going to figure it out.”
Comments, questions, or feedback? Email us at feedback@voguebusiness.com.
More from this author:
How Printemps won over New York
