For decades, online shopping has hinged on one frustrating constant: the search bar. To find what you want, you’ve had to guess the right keywords, sift through irrelevant results and hope that a brand’s backend taxonomy aligns with how you describe a product. But now, with generative AI tools like ChatGPT and a wave of fashion-focused search startups, the model of product discovery is undergoing a rapid transformation.
So how are people really searching today? More than 60% of Gen-Z shoppers now start their product searches outside of Google or Amazon, according to a 2024 study from Forbes. Instead, they’re browsing TikTok hauls, using Google Lens or Pinterest boards, or leaning on private chats or resale platforms. They want discovery based on vibe, value and social context, not rigid dropdown filters. That shift has made traditional e-commerce search tools feel not only outdated, but broken.
Phia, which launched April 24, is one of the clearest examples of what the next wave of fashion search looks like. Co-founded by Phoebe Gates and Sophia Kianni, the browser extension and app works as a digital shopping assistant for Gen Z, comparing prices of new and secondhand products across more than 40,000 marketplaces.
“We wanted something that took the guesswork out of smart shopping, especially when you’re navigating between full-price retail, secondhand, and dupes,” said Kianni.
The tool’s “Should I Buy This?” feature helps users determine if they’re overpaying, factoring in product data, market history and resale listings. “It’s about making it easy to understand what something is actually worth, not just what it costs,” Gates said.
Phia surfaces resale listings in real time from platforms like The RealReal, Poshmark and Vestiaire Collective, offering what Kianni calls “invisible value” while you browse. The startup has raised $850,000 to date, with investors including Kris Jenner, Sara Blakely and Joanne Bradford.
“People are interested in a vibe, they’re interested in advice,” said Julie Bornstein, co-founder of the AI shopping startup Daydream. “You shouldn’t need to know the exact right term to find a black block heel sandal.”
Daydream, set to launch this June, enables search by natural language and images. Shoppers can ask things like, “What should I wear to a summer wedding in Costa Rica?” or upload an image and say, “I want this in blue.” The platform has already onboarded over 2,000 brands and retailers, including Net-A-Porter, Jimmy Choo and Alo Yoga.
Unlike generalized platforms like Lyst or Google Shopping, Daydream’s strength is in vertical specificity and context. The startup raised $50 million in seed funding from Forerunner Ventures, Index Ventures, Google Ventures and True Ventures.
“We’ve trained consumers for the last 20 years to guess the right taxonomic term,” Bornstein said. “Now we can use language and images together in a way that allows people to simply express what they need.”
That kind of contextual search is quickly becoming table stakes. “Even if you create the best shopping app in the world, your customers are still going to Amazon and Google,” said Jess Meher, svp of marketing at Loop Returns. “That’s why it’s critical that brands start thinking about how their product data shows up in AI ecosystems.”
Meher said, due to emerging search capabilities, most retailers will need to rethink how they structure everything from image galleries and copy to shipping speeds and return policies. “These are all decisions consumers make in milliseconds,” she said. “AI can surface or bury a product based on those signals.”
Meanwhile, OpenAI is building its own vision of the future of shopping, inside ChatGPT itself. On April 23, lines of code uncovered by TestingCatalog revealed that OpenAI was preparing native Shopify support. The update, now live, includes embedded price and shipping fields, images, reviews, and even direct checkout buttons.
As of May 13, ChatGPT’s latest GPT-4o model now offers real-time product suggestions in categories like fashion, beauty and home goods. Users can ask detailed shopping queries in natural language and receive customized results, complete with direct links. OpenAI says the recommendations are powered by third-party metadata and not ad placements. Currently, it does not take affiliate fees, but CEO Sam Altman has indicated that could change.
ChatGPT now fields more than a billion web searches per week. And with updates like memory integration and a WhatsApp extension on the way, OpenAI is building a competitive layer atop traditional search engines. TestingCatalog describes it as turning ChatGPT into a full-fledged e-commerce platform.
