Luxury fashion brands face a recurring challenge. Every season introduces new products, making historical SKU level comparisons impossible. This project shows how AI can support more reliable sales forecasts by shifting the focus from individual items to statistically meaningful product categories.

An Italian luxury fashion brand operating across multiple business units, each responsible for ordering products for new seasonal collections.

The brand needed to make informed ordering decisions without the ability to rely on repeated SKUs across seasons.

Different business units followed distinct decision logics, often combining experience with non-data driven criteria. As a result, forecasting processes were manual, time consuming, and prone to inaccuracy.

Advanced analytics and AI based forecasting techniques applied to custom product categorization, enabling aggregation and comparison across seasons.

The project introduced a new forecasting logic based on statistically meaningful product groups.

Products were aggregated into custom categories derived from available attributes. Attribute values were clustered into coherent groups, such as combining multiple shades into a single color category.

Both historical and upcoming products were mapped to these categories, enabling sales forecasts at category level for each business unit.

  • Data-driven support for ordering decisions
  • Reduced effort compared to manual forecasting approaches
  • More consistent and harmonized processes across business units

The initiative was developed as a pilot project and requires further testing, refinement, and automation before full operational rollout.