Published
3 min read
By Tom Whitfield

Ecommerce & Digital

J. Crew Powers New Style Advisor With Anthropic's Claude Models

J. Crew launched Style Advisor on JCrew.com, powered by Anthropic's Claude. Its in-house team prototyped with Claude Code and Claude Opus 5.5 before shipping production.

J. Crew turns to Claude to create new Style Advisor tool
PhotoJ. Crew turns to Claude to create new Style Advisor toolAI-generated

At a glance

  • J. Crew launched Style Advisor, a Claude-powered styling assistant, announced Nov. 12
  • The in-house team prototyped with Claude Code and Claude Opus 5.5, then shipped production on Claude Fable 5.1
  • Claude Sonnet 5.5 handles styling feedback; Claude Haiku 5.5 handles shorter tasks
  • J. Crew ranks No. 80 in the Top 1000 Database and No. 135 in the AI Commerce Rankings

J. Crew has put an AI styling assistant built on Anthropic's Claude live on JCrew.com, the two companies announced in a Nov. 12 post on the Claude website. The tool, called Style Advisor, delivers personalized recommendations to online shoppers and runs entirely on the retailer's own site rather than a third-party app.

The launch matters for practitioners because of who built it: J.Crew's in-house product team, not an outside vendor. According to Anthropic, the team first created a prototype using Claude Code and Claude Opus 5.5, then its engineers carried that prototype into production using the Claude Fable 5.1 model, which Anthropic markets for "complex, open-ended work."

What does the assistant actually do for shoppers?

Style Advisor is accessible from a menu option on JCrew.com and opens inside the web browser. The interface displays suggested prompts so shoppers can start a conversation without typing an open-ended query.

Behind the scenes, two models split the workload. The Claude Sonnet 5.5 model receives prompts and returns styling-focused feedback, while the Claude Haiku 5.5 model handles shorter, more focused tasks. The architecture mirrors a pattern trade readers will recognize: a frontier model built the experience, faster and cheaper models run it.

J.Crew positions the output as stylist work, not algorithmic guessing. The company maintains that because the styles being referenced all come from its human team, the recommendations are designed to stay consistent with the brand's priorities. The tool is meant to offer recommendations in the way a J. Crew stylist would when responding to a customer in person.

Why J. Crew chose to build rather than buy

CEO of J.Crew Group Libby Wadle framed the move as a deliberate attempt to get ahead of changing shopping behavior.

"As consumer expectations evolve and AI opens up new possibilities for retail, J.Crew is challenging the traditional model of online shopping, moving beyond individual product searches toward a more connected and inspiration-led experience," Wadle said. "For a long time, we've been thinking about how customers will want to shop in the future, and how we can lead that evolution rather than react to it."

She extended the point to the brand's merchandising depth. "We see an opportunity to fundamentally rethink the shopping experience," Wadle said. "Through our work with Anthropic, we're bringing decades of J.Crew styling, merchandising and creative expertise together to help shape what's next for customer behavior."

Paul Smith, chief commercial officer at Anthropic, credited the retailer's in-house team with the execution. "J.Crew has one of the most recognizable points of view in fashion, and its team used Claude to bring that expertise to customers in a new way," Smith said. "They used a frontier model to build the experience and faster, more efficient models to run it, so customers get quick answers that stay true to the brand's styling."

Where J. Crew stands in the rankings

J. Crew ranks No. 80 in the Top 1000 Database, which tracks North America's largest online retailers by annual ecommerce sales. The retailer also holds the No. 135 position in Digital Commerce 360 and ReFiBuy's AI Commerce Rankings, a set of rankings available through the Top 1000 Pro with AI Database.

For marketers at mid-sized and larger retailers, the deployment sets a workable template: prototype with a frontier model through developer tools, ship production on the same model family, and keep brand voice intact by grounding answers in the internal styling team's catalog. The question other apparel marketers now face is whether they can replicate that in-house build path before their competitors do.

Original: subscription.digitalcommerce360.com

Share this article:

More from Tom Whitfield

Tom Whitfield

Show full bio

Staff writer covering people, launches and day-to-day practice in trade marketing, shopper marketing, trade promotions, category management at Target Marketing.

138 articles

Related articles

October 12, 2026Kroger, Giant Eagle and DoorDash Bet on Agentic AI Shopping Assistants
October 30, 2026Fanatics’ OpenAI tool lifts online outfit engagement 10%
September 30, 2026Amazon Opens Seller Data to Claude in Agentic AI Push
September 28, 2026Groceryshop 2025: AI Talk Turns to Results, Value Gets Redefined
October 30, 2026Etsy Posts 8.8% GMS Jump and Draws Lines Around AI Use

« Previous article