Fashion Retail Competitive Intelligence & GenAI Trend Platform

Context

LC Waikiki is a fashion retailer with 1,000+ stores in 50+ countries plus e-commerce. Its merchandising and pricing teams need to see what competitors are doing without waiting on manual research. This production platform was built for them, tuned to the company’s own brand and category mix.

Role: designed and built the platform, as Data Solution Architect.

Problem

Pricing and assortment decisions took a week, and competitor research was manual. A comparison across brands also needs every product in one shared taxonomy, whatever each brand calls it.

Approach

• Collection: Python scrapers on Cloud Run, which scale with load; agents discover and validate data sources.
• Normalization: an LLM classifier maps products to a shared taxonomy; LLMs also summarize trends.
• Benchmarking: price, stock and colour are compared across 20+ brands and ~95 markets.
• Trend signals: 200 social accounts are monitored; colour and category summaries go straight to business users.

Architecture and technology

Python scrapers on Cloud Run; Delta Lake on Databricks for raw and curated competitive data; LLMs for product normalization and trend summarization; BigQuery for analytics and merchandising dashboards.

Outcome

• Pricing and assortment decisions: from 1 week to 1 day.
• Manual competitor research replaced by continuous, automated collection.
• Coverage: 20+ brands, ~95 markets and 200 social accounts.

Lessons

• The gain in decision time rests on comparable data. The LLM’s main job here is to place products in a shared taxonomy: a data-preparation step, not the decision itself.
• Collection, storage and serving are separate layers: raw and curated data in Delta Lake, analytics and dashboards in BigQuery.