J. Crew launches AI Style Advisor to mimic human fashion experts

Moving beyond simple search

For forty years, J. Crew has relied on human stylists to curate its preppy aesthetic of cashmere cardigans and tweed jackets. While the brand has grown from a catalog into a major omnichannel retailer, the creative decisions regarding how to pair accessories and clothing were previously made by people. Now, the company is introducing artificial intelligence to handle these styling decisions online.

The new tool, called Style Advisor, is designed to move past simple product discovery. Instead of just showing a customer the specific item they clicked on, the AI aims to suggest a complete look. As Danielle Schmelkin, chief information officer at J. Crew, explained, the goal is to replicate what stylists bring to the table by recommending more than just a single piece of clothing.

This shift comes as more people move toward agentic shopping, which involves using AI agents to navigate the purchasing process. According to consulting firm McKinsey, while many people use general AI tools like ChatGPT, only 27% of consumers are currently using AI assistants created specifically by retailers.

How the AI learns style

To ensure the tool reflects the brand's specific identity, J. Crew trained the AI on a massive dataset. This includes thousands of outfit pairings from the company's professional photo shoots. The system also incorporates J. Crew’s established styling principles and specific rules for pairing accessories with various garments.

The technology is powered by a variety of models from Anthropic. The company used Claude Opus 5.5 to build a working prototype of the tool. For the live experience, they use Fable 5.1 to handle complex and ambiguous judgments, while Claude’s Sonnet 5.5 and Haiku 4.5 provide the speed needed for fast back-and-forth conversations with shoppers.

The AI is grounded in contextual details, meaning it can process specific nuances provided by the user. If a customer mentions they are attending a wedding in Miami but want something less formal, the tool can adjust its suggestions accordingly. It is designed to understand the intent behind the request rather than just matching keywords.

Personalization and scale

In the future, J. Crew plans to make the Style Advisor even more personal by pulling in a consumer’s past purchase history. This will allow the tool to replicate the omnichannel experience where a store associate might already know a customer's preferences. By combining historical data with real-time questions, the AI can offer highly tailored advice.

Because running these models can be expensive, J. Crew built a back-end observation layer to monitor costs. Schmelkin noted that the company wants to ensure they are using the models efficiently to make the tool scalable. They are constantly refining the outputs to balance high-quality customer service with the technical costs of running the AI.

The tool will first appear on the company's desktop website before moving to the mobile app. This rollout allows the brand to test the technology with a wide audience while refining the user experience. The goal is to provide a seamless transition from a digital search to a curated fashion recommendation.

Internal AI adoption

The creation of Style Advisor was not a top-down mandate from the company's executive suite. Instead, it was conceptualized by a digital product designer who developed a prototype before engineers were brought in. This approach reflects Schmelkin's effort to encourage employees to find their own ways to use and build with AI technology.

To support this culture, J. Crew has launched an AI fluency program for its staff. This training starts with basic concepts and moves into more advanced techniques like vibe coding. The company also established an internal AI lab called Second Thread, which has expanded from a tech-only group to include all corporate employees.

Employees are also encouraged to use enterprise tools from both Anthropic and OpenAI in their daily work. J. Crew has also hosted a two-day tech summit for four years, featuring workshops and hackathons. These initiatives are designed to help the workforce build a habit of using AI to improve team productivity and solve problems.