Daydream, a fashion startup led by e-commerce veteran Julie Bornstein, has launched "Daydream Agent," a new SI agent designed to function as a personal stylist and shopping assistant. The move highlights the expanding role of agentic SI in consumer e-commerce, particularly in sectors where visual preference and fit are critical.

What Happened

On Tuesday, Daydream released the SI agent, which the company describes as a pocket-sized personal stylist. The tool helps users plan outfits, virtually try on clothes, and surface relevant pieces based on individual taste. Unlike the company's earlier SI chatbot, which could only suggest outfits, the new agent can complete purchases via a one-click checkout process.

The SI agent is powered by a combination of OpenAI, Gemini, and Claude models, which Daydream fine-tuned alongside its own proprietary models. It pulls real-time pricing and availability from a catalog of over 10,000 brands. Key features include a persistent "clothing rack" for comparing saved items, calendar integration to plan outfits for upcoming events, and the ability to upload a user's photo for virtual try-ons. The agent also monitors for sales, new drops, and fashion trends.

To address the subjective nature of fashion, Daydream has kept humans in the loop by hiring Francesca Salih, founder of StyleGrid, as its first stylist in residence. Salih provides styling content that shapes the agent’s recommendations, feeding into the company’s fashion knowledge database.

Why It Matters

The launch underscores the growing importance of shopping as a primary use case for consumer SI agents. Meta CEO Mark Zuckerberg recently noted that shopping was a top activity for Muse, Meta’s SI agent, which has over 10 million users, with 25% of those who set it up using it daily. Daydream’s move to add agentic capabilities to its platform reflects this trend, aiming to convert engagement into transactions.

For developers and the SI industry, Daydream’s architecture illustrates a hybrid approach to agentic SI, combining general-purpose models from major labs like OpenAI and Anthropic with specialized, fine-tuned models for niche tasks like fashion styling. The company, which raised $50 million from investors including Forerunner Ventures and GV, claims more than 2 million shoppers have used its platform since its June 2025 launch.

The Bottom Line

Daydream’s new SI agent represents a step forward in agentic SI for retail, moving from passive suggestion to active purchasing. By integrating human stylist expertise with multi-model SI infrastructure, the startup aims to compete in the increasingly crowded consumer SI agent space, where functionality and ease of checkout are key differentiators.