Phia raises $35M for agent-led AI shopping

Agent-led commerce and the “missing layer”

Online shopping still asks people to do too much work: bounce between tabs, decode inconsistent sizing, ignore aggressive retargeting, and sift through filters that rarely capture real intent. In practice, that friction hurts everyone—shoppers waste time and lose confidence, while brands face higher return rates and weaker loyalty because the discovery journey is noisy and impersonal.

That’s the opening Phia is betting on: an “agent-led” shopping experience where software sits between people and products and actively helps them decide, not just browse. For the ai world organisation, this shift matters because it signals a broader transition toward AI systems that act, reason, and personalize in real time—exactly the kind of applied, business-facing innovation we spotlight through ai world organisation events and discussions at the ai world summit.

In other words, Phia isn’t trying to be another storefront or another feed. It’s trying to become the next decision layer for commerce—where “what I’m trying to buy” is interpreted instantly, “what I’ll actually like” is modeled as taste, and “what I should trust” is supported through transparent, in-context recommendations.

Phia’s $35M raise and rapid early scale

Phia, launched in April 2025 by co-founders Phoebe Gates and Sophia Kianni, announced a $35 million Series A at a $185 million valuation. The round was led by Notable Capital, with participation from Khosla Ventures and returning investor Kleiner Perkins.

In just 10 months since launch, Phia says it surpassed 1 million users and partnered with more than 6,200 retail brands, spanning contemporary labels through luxury fashion houses. The company also reported that it drives millions of dollars in sales for brands each month and has achieved 11x revenue growth since launch.

Phia also attributes much of its user growth to organic, founder-led marketing, reporting over 2 million followers and 430+ million views across its owned platforms. That kind of distribution-first execution is increasingly a pattern in consumer AI: ship a useful workflow, build trust with a community, then scale the underlying intelligence once retention and repeat behavior become visible.

For teams tracking consumer AI at the ai world summit 2025 / 2026, the headline isn’t only the money—it’s the speed at which a new shopping behavior can form when an AI agent reduces effort and uncertainty.

How Phia’s AI agent positions “intent, taste, trust”

Phia describes itself as building an “AI alignment layer” between brands and consumers. Practically, that means embedding an AI shopping agent directly into a user’s natural shopping flow and offering real-time insights that help people make better purchase decisions while they browse.

On the infrastructure side, Phia says its agent ingests billions of new products and processes millions of searches daily. The company also reports it has already reduced search latency by 80% and increased monetized GMV by 40%, positioning performance and speed as core differentiators rather than “AI” as a vague feature.

What’s especially notable for anyone attending ai conferences by ai world is the product philosophy implied here: in the next era of commerce, “search” is not a destination page—it’s an ongoing, contextual assist that follows the user. Phia’s roadmap points further in that direction, stating that the next phase will include individualized rewards, taste-aware recommendations, and community-curated digital closets.

Phia also says it plans to evolve toward real-time LLM agents that personalize each user’s shopping experience, including intelligence designed to predict consumer preferences. If that trajectory holds, the shopping journey becomes less like navigating a catalog and more like consulting a tailored advisor that learns from your actions while staying embedded in the moment of decision.

What brands get: performance model, dashboards, and measurable outcomes

A central part of Phia’s pitch is its “zero-dollar upfront, performance-based model,” positioning it as low-friction for brands to try. The company reports that brands on Phia see 13% higher conversion rates, 30% stronger new customer acquisition, 15% increased average order value, and return rates reduced by more than 50%.

Those numbers matter because returns and acquisition costs are two of the biggest pressures in modern e-commerce, and they’re also areas where “generic personalization” often fails. If an agent can genuinely interpret intent (what the shopper is trying to accomplish), map that to taste (what they’ll like and keep), and support trust (confidence that reduces second-guessing), it can improve conversion and reduce returns at the same time—an unusually powerful combination.

Phia says it will expand its partner offering with dashboards that provide real-time visibility into audience behavior, emerging trends, and category positioning, among other insights. This is where the product stops being “a consumer app” and starts becoming a two-sided intelligence layer: shoppers get assistance, while brands get clearer signals about demand and decision drivers.

At the ai world organisation, we look at this as a concrete example of how AI is moving beyond content generation into operational leverage—optimizing funnels, reducing waste, and making discovery more efficient. That’s also why themes like agentic workflows, personalization, and responsible data use consistently show up in sessions and networking conversations at the ai world summit and across ai world organisation events.

What’s next, and why it’s relevant to AI World Summit 2025 / 2026

Phia says the new funding will be used to accelerate its core AI capabilities, build out real-time agents, and grow the team needed to execute quickly. It also states that it will expand its network of brand partners as it evolves into an end-to-end shopping agent and discovery platform for the next generation of buyers and brands.

From a market perspective, Phia’s story highlights a broader trend: consumer experiences are being rebuilt around “agent interfaces,” not just better UI. When AI can act as an interpretation layer—turning messy human intent into structured decisions—it changes the economics of discovery and reshapes where value accrues in the shopping journey.

For readers following the ai world summit 2025 / 2026 agenda themes and the wider ecosystem of ai conferences by ai world, the takeaway is straightforward: agent-led models are no longer a concept demo; they are raising meaningful capital, shipping into real workflows, and reporting measurable outcomes. If you’re a brand, retailer, marketplace, or marketing leader, it’s a signal to evaluate where “AI agents” can reduce friction without sacrificing trust—and to think early about the governance, attribution, and customer experience standards that will define the next layer of commerce.

As the ai world organisation continues building global conversations through the ai world summit and other ai world organisation events, stories like Phia help ground the hype cycle in real adoption metrics and practical business impact. That grounding is where strategy becomes actionable: understanding which agent experiences create repeat behavior, which incentives reduce returns, and how brands can participate without losing their direct relationship with the customer.