J.Crew has introduced Style Advisor, an AI-driven styling service accessible on its American website, powered by Anthropic's Claude technology. The tool became available this week for both women's and men's clothing categories across the US market.
Rather than requiring customers to hunt separately for individual garments, the system allows shoppers to describe their needs—whether dressing for a beach wedding or preparing for a first day at a new job—and receive curated complete looks in response. Users can then refine these suggestions iteratively or begin with a single piece and request complementary items.
The distinguishing feature of Style Advisor is that every recommendation draws exclusively from outfits assembled by J.Crew's in-house styling team during the brand's catalogue photography sessions. Rather than generating novel combinations, the AI retrieves and presents these pre-existing looks, ensuring recommendations stay true to the retailer's aesthetic vision.
"We see an opportunity to fundamentally rethink the shopping experience," said Libby Wadle, CEO of J.Crew Group.
The implementation relies on multiple Claude models working in tandem. Claude Sonnet 5.5, the mid-tier variant, functions as the primary stylist during conversations, while Claude Haiku 5.5, the smaller model, manages routine operations to ensure rapid response times. Currently, the tool does not customize suggestions based on individual purchase history, though J.Crew intends to incorporate this capability and expand the service internationally in the future.
The fashion retail sector is witnessing accelerating adoption of AI styling solutions. Ralph Lauren rolled out its Ask Ralph assistant during the previous year, Kate Spade integrates Amazon's shopping AI, and Zalando has announced plans to launch an agent capable of shopping across competing retailers beginning in 2027. Data from Prime Day shopping events demonstrated that customers directed by AI assistants exhibited higher purchase conversion rates compared to those arriving through conventional search methods.
Source: The Next Web
