AI Retail & eCommerce Agent

Description

We run a fashion storefront with about 8,000 SKUs and a return rate of 31%, most of it sizing. Our product pages carry a size chart nobody reads and a fit description written by whoever loaded the product. The job is a shopping assistant focused on fit and suitability: it should use the customer's purchase and return history, what they kept and what went back, and the actual measurements in our PIM, to answer "will this fit me" better than a static chart does. It must be willing to say "this brand runs small, consider the next size" and equally willing to say "based on what you returned last time, this probably will not work". Reducing returns is worth more to us than an extra order.

• Build fit guidance from PIM measurements, brand-level fit patterns and the customer's own return history
• Give honest discouraging answers where the data supports them, not just upsells
• Handle logged-out shoppers with a reduced but still useful experience
• Integrate on product pages and in the cart, in the storefront's own design language
• Measure return rate on assisted versus unassisted purchases

• Fit assistant on product and cart pages
• Fit model combining PIM data, brand patterns and customer history
• Logged-out experience
• Return-rate measurement with an assisted/unassisted comparison