Phia: the AI shopping agent that grew in public
Phia: the AI shopping agent that grew in public
Style
Fashion
Agent
Social Media
Date published:

Style · Fashion · AI commerce
Two 23-year-olds raised $43M and built a million-user AI shopping agent — and the most transferable lesson isn't the tech. It's that they turned their own growth metrics into the marketing.
The setup
Phia is an AI shopping assistant. You're about to buy something online, you tap "Should I buy this?", and Phia checks that item's price across 40,000+ retail and resale sites — new and secondhand — then tells you whether the price is fair and where the same thing is cheaper. Under the hood: a database of 250 million-plus secondhand listings and a model that matches an item by image, brand, and attributes. It launched in April 2025, founded by Phoebe Gates and Sophia Kianni, Stanford roommates and co-CEOs.
The numbers
Inside its first year: 1 million-plus users and $43M+ raised at a $185M valuation — an $8M seed (Kleiner Perkins) followed by a $35M Series A led by Notable Capital, with Khosla in the mix and an angel list that reads like a culture pass: Kris Jenner, Hailey Bieber, Sara Blakely. Pointedly, they didn't take money from Phoebe's father.
The real lesson: they built in public
The tech is good, but AI shopping is crowded — OpenAI, Walmart, and a dozen startups are circling. Phia's edge was distribution, and specifically the decision to build in public. They run a podcast (The Burnouts) with hundreds of thousands of followers, publish their experiments, ask users to DM feedback, and — the sharp part for anyone raising — post their user and revenue growth openly on LinkedIn. Investors reached out to them because the traction was visible. Their CAC is near zero because the founders themselves are the acquisition channel.
When it came time to close the Series A, the founders were explicit: conversations often started from that public content, but "data is king" closed the round. Visible distribution power (hundreds of millions of content views), near-zero CAC, and a clear revenue model — laid out in decks and Notion boards — is what turned warm relationships into term sheets.
The investable insight
Two things. First, the wedge is the agent. Phia's real ambition isn't price-checking; it's becoming the layer that decides what a generation of women buy — "the agent that determines what women buy," in Gates' words. Whoever owns that decision owns the most valuable real estate in commerce. Second, and more immediately useful here: in a crowded AI category, distribution and openness are the differentiator. Founders who share their metrics loudly compound an advantage the quiet ones never get — inbound investors, inbound talent, inbound press, all at zero marginal cost.
For the room
Building in public is a fundraising strategy. Post the growth. Visible traction pulls investors, talent, and press to you. Silence compounds nothing.
Be your own distribution. When the founders are compelling, founder-led content beats paid acquisition and drives CAC toward zero.
Own the decision layer. In AI commerce, the durable position isn't the store — it's the agent that decides what people buy.
The takeaway: In a race where everyone has similar AI, Phia won attention by being loud about the right things: real users, real revenue, in public. If you're building and raising, that's the playbook — your traction is your best pitch, so show it.
