A new wave of retail innovation is taking hold at Phia, the AI-driven shopping startup co-founded by Phoebe Gates and Sophia Kianni, which launched in April 2025. With its free mobile app and browser extension, Phia is designed to bring Gen Z-style convenience, sustainability and personalization into the world of online fashion shopping.
The vision behind the startup
Gates and Kianni met as roommates at Stanford University and identified a key friction point: “Consumers still waste hours comparing prices and hunting for deals, only to still end up overpaying,” Gates said. (PR Newswire) They therefore built Phia to act like a shopping assistant that aggregates data across new and second-hand items to deliver smarter buying decisions.
“There felt like this giant white space … what should we actually buy, and why doesn’t everyone have a personal shopper in their pocket?” Kianni said. (Yahoo Tech)
How AI is powering the transformation
Phia pulls together massive datasets—covering billions of fashion products—including more than 300 million secondhand items in one of the U.S.’s largest resale databases. (Digital Commerce 360)
On a user’s device, when browsing an item, Phia can instantly answer: “Is this a fair price?” Or suggest cheaper alternatives from tens of thousands of retailers. (PR Newswire)
The startup is developing a personalized AI shopping agent—a proprietary large-language-model (LLM) trained on user transaction and browsing data—to recommend what to buy, when to buy it, and what resale value it may retain. (PR Newswire)
Sustainability and resale are embedded: Phia’s co-founders highlight that secondhand shopping can reduce carbon footprint by roughly 80% compared with new items. (Yahoo Tech)
Key features that reflect this shift
A “Should I Buy This?” browser-button, which triggers instant AI-driven price comparison across retail and resale sites. (The Times of India)
Recommendations on resale value: For instance, if you’re considering a $500 handbag, Phia may show you its current resale value ($300–$400) or advise you to buy a cheaper alternative if the expected depreciation is high. (Yahoo Tech)
Brand and item discovery from a large partnership network (over 5,000 brand partners in early months) enabling more options and better deal visibility. (Digital Commerce 360)
Why this could matter
The founders argue that online shopping is overdue for disruption. Most e-commerce tools help with coupons or one site; Phia’s goal is to unify search, pricing, resale and AI-recommendation into one workflow.
For consumers: it promises less time wasted, better deals, and more informed purchases.
For brands and retailers: it may shift how customers discover, evaluate and buy items—potentially reducing acquisition costs and improving targeting. (Digital Commerce 360)
Obstacles & questions ahead
Scaling the AI-agent: Building a personalized model that works across hundreds of millions of items and millions of users is expensive and complex. Phia is investing in GPU infrastructure and proprietary models to keep cost and latency manageable. (PR Newswire)
Data and privacy: Using transaction, browsing and resale data raises questions about user consent and transparency over how “personal shopper” recommendations are generated.
Market competition: Many tools compare prices or aggregate deals, but Phia’s uniqueness will depend on seamless user experience and measurable savings.
Resale market dynamics: The resale value of items can vary widely by brand, region and condition. Accurate prediction remains a challenge.
