Retail veteran Julie Bornstein launches Daydream AI Fashion Shopping Agent

“I remember one day I wanted to build a search and discovery platform that would help you find the right one,” she said. luck In an interview earlier this month.

In the role of executives at Nordstrom over the next few decades, Urban Outfitter, Sephoraand Stitch Fix, and as the founder of fashion startups, yes, she Pinterest In 2022, Bornstein has come close to his goal.

But on Wednesday, Bornstein, 52, announced what she considers to be the most complete manifestation of her quarter-century vision to date. that Available in the form of AI-powered fashion discovery and shopping marketplaces Fantasywill begin in beta a year after Bornstein first announced its $50 million seed investment.

“This is the culmination of everything I saw and did,” Bornstein said. “We were waiting for technology to catch up.”

For Bornstein, the first seed for the Daydream idea In 2022, CHATGPT entered the scene and began training queries that said they could create subtle, detailed answers and recommendations with a language model for conversation or LLMS.

Now, Bornstein’s Daydream is trying to beat many existing fashion brands and retailers, and take down Anch-Me-Anything chatbots like ChatGpt and Prplexity. Anyone who splits this first can use the benefits of first movers to build a highly engaged, large consumer beat-hard flywheel that attracts partner brands and retailers who don’t want to miss out on new growth channels.

At launch, customers moving to Daydream.ing (there is no app yet) will first come a series of biographical questions (date of birth, gender) followed by subsequent sizing, favorite brands, and budget questions (the fact that the cheapest hurd budget range ranges up to $150 speaks about customers about Daydream customers. Events, moods, or products you are shopping for today?” (New users can skip a short introductory survey if necessary, and skip ahead.)

This question comes with a large box to type. This appears to be intentionally designed to derive longer queries than simple keyword searches. In a simple test of this reporter, the system gave some queries fairly good results. “In addition to looking for sneakers that can be worn for kids’ sports games, you can also get away even at moderately flashy dinners.”

Once you enter one query, the system immediately generates a flood of product images, each appearing on a rectangular tile, taking over two-thirds of the page.

The Daydream agent (or chat experience) moves to the left of the screen and asks them to try and promote feedback to further refine their results (users can tap on the product they like and ask for something similar at a different price.

To purchase items, click from Daydream to the retailer or brand’s own website. (In the future, Bornstein plans to close the transaction on its partner’s shopping site on behalf of its customers.) The fact that the transaction will take place on the brand and retailer’s site rather than within the Daydream is one of the reasons why Bornstein believes that so many brands and multi-brand retailers have partnered from launch.

“You could see us as a threat to them, but that’s not,” Bornstein said. “Everyone is so dependent on me I’m so tired of it. Google Facebook for new users. ”

Daydream has signed up with around 200 partners so far, most of which are apparel brands. Nike It lines up in Madewell and from luxury labels like Chloe to Khaite, it has around 15 multi-brand retailers, including Nordstrom and Net-A-Porter. Bornstein does not offer any specific reductions, but the startup will charge a referral fee if Daydream users continue to buy from them. She allowed their rates to be higher than affiliate network referral fees (range 5% to 10% in fashion), but lower than some of the fees charged in the luxury market, which is over 30%.

These partner brands and retailers provide product catalog data to startups. This catalog data allows Daydream to enrich additional attributes, allowing AI agents to create quality recommendations in response to nuanced questions. At launch, Daydream results are influenced by approximately 12 small language models, each dealing with specific attributes such as color and fabric.

Bornstein believes that all AI chat and search experiences are like ChatGpt, Perplexity, and Gemini, but are extremely popular in product-related searches in categories such as TVs and hair dryers. However, she also believes that the next breakthrough will not be clear anytime soon, as the generative, agent AI impact on shopping is moving so fast.

“I think it’s hard to imagine how e-commerce can evolve,” Bornstein said. “What we built today is a kind of bridge between where it was and where it was heading.”