J.Crew Style Advisor: an AI stylist built on Claude
J.Crew's AI Style Advisor runs on Anthropic's Claude models, grounded in its own outfit pairings, with cost monitoring built in. Here's the playbook for your store.
J.Crew did not build a chatbot that searches its catalog. It built an AI shopping assistant that styles you. On November 11, the retailer launched Style Advisor, Fortune reported. It runs on a mix of Anthropic's Claude models and is grounded in J.Crew's own styling data. Three choices make it worth studying: the data it learns from, how it splits work across models, and the fact that cost tracking shipped on day one.
What actually happened
Per Fortune:
- What it does. Style Advisor asks questions to find intent, like a wedding in Miami that is "not too formal." Then it recommends pieces and how to wear them together. CIO Danielle Schmelkin: "It's not just about the piece that you're looking for, but how to style it."
- The data. It is trained on thousands of outfit pairings from J.Crew's photo shoots, plus the brand's styling principles and accessory pairings.
- The models. A top-tier Claude model handles the ambiguous judgment calls. Faster Sonnet- and Haiku-class models handle the quick back-and-forth. J.Crew's team used Claude Opus to build the first working prototype.
- Cost control. J.Crew built "a back-end observation layer" to monitor costs closely, Schmelkin said, so the tool can scale.
- Where. Desktop site first. The mobile app comes later. Past purchase history will feed recommendations in a future update.
J.Crew shared no conversion or usage numbers.
Why it matters for your business
You do not need J.Crew's budget to copy J.Crew's playbook. The expensive part is not the model. It is the data only you have.
J.Crew's edge is thousands of outfit pairings that no competitor has. Your version might be your best salesperson's answers to the ten questions customers always ask, your bundles, or your fit notes. A generic model with your catalog gives generic answers. A model grounded in your know-how sells like your best employee.
The routing matters too. Sending every "do you have this in blue?" to the biggest model burns money. Send hard judgment calls to the big model and the rest to the small one. Then watch the cost per chat from day one, like J.Crew did, not after the first invoice.
Start small:
- Write down your styling or buying rules before you write a prompt.
- Route by difficulty: a small model for lookups, a big one for advice.
- Log the cost per conversation and compare it to the order value.
Key takeaways
- J.Crew launched Style Advisor, an AI stylist on Anthropic's Claude models, on its desktop site
- It is grounded in thousands of in-house outfit pairings and styling rules
- Bigger models handle judgment; faster models handle quick replies
- Cost monitoring was built in from launch
- No performance numbers yet; mobile and purchase-history personalization come later
Want a shopping assistant that sounds like your best salesperson, not a search box? We build AI assistants grounded in your own product knowledge, with model routing and per-chat cost tracking. Run the numbers on our ROI calculator or tell us what your customers always ask.
Sources: Fortune.
- #ai-shopping-assistant
- #jcrew
- #anthropic
- #claude
- #ecommerce
Tommy Rush — Founder, Rush Commerce
Operator turned builder. 15+ years running operations — now shipping the systems businesses run on. More
Get The Rush Report weekly — one email, zero fluff.
Keep reading
Square Squaremates goes GA: an AI agent in your Dashboard
Square's Squaremates AI business agent is now GA for U.S. sellers at no added cost, with a Square CLI and an MCP that can make changes. Set approvals first.
Read itUnited Airlines bans AI agents from booking flights
United's new contract of carriage bars AI agents from booking or changing tickets, even with customer approval. Decide your own agent checkout policy now.
Read it