A procurement consultant needed to regularly compare pricing and specs for industrial equipment across 8 different supplier websites — a process that was taking a junior analyst 2 days per report cycle.

How It Works

The tool combines a Python scraping layer with an intelligent data normalization engine:

  1. Scraping: Playwright-based scrapers extract product listings, prices, specs, and availability from each supplier site on a scheduled basis
  2. Normalization: A custom NLP pipeline maps different spec names to a unified schema (e.g., "Max Load" = "Maximum Capacity" = "Load Rating")
  3. Comparison engine: Products are matched across sites by model number, EAN, or fuzzy name matching
  4. Report generation: Automated Excel reports with conditional formatting (green = lowest price) and PDF summaries using ReportLab

Results

Report generation time: from 2 days → 45 minutes (fully automated). The consultant now runs fresh comparisons daily and has found consistent savings of 8–15% by catching price discrepancies across suppliers.