A web-scale price intelligence platform that finds the cheapest item across stores so customers pay the lowest receipt-level price.
- Coverage
- Multi-retailer
- SKU matching
- Cross-store
- Refresh
- On a schedule
What we shipped
Hercules guarantees the lowest total receipt for FMCG groceries by buying each item at its cheapest source. We built the data platform that scrapes retailer websites continuously, normalizes product data, and matches SKUs across stores so the dispatch team always knows where every item is cheapest right now.
Hercules built a business model on a hard guarantee: the customer pays the lowest possible total on the receipt, even when that means physically buying the bread at one store and the yoghurt at another. That promise only works if the price intelligence behind it is comprehensive, current, and accurate.
Blueprint → AI Pilot → Production launch → Scale and operate.
We followed the Datablooz Delivery Model. See our process.
- Blueprint
Mapped target retailers, product taxonomies, refresh requirements, and the dispatch workflow that consumes the data downstream.
- AI Pilot
Built and validated web extraction agents and SKU-matching logic on a subset of FMCG retailers, benchmarking match accuracy against manual review.
- Production launch
Deployed scheduled scraping pipelines, normalization, cross-store SKU matching, and a structured price-intelligence dataset feeding the Hercules dispatch and ordering system.
- Scale and operate
Added price-trend monitoring, promotion detection, and similarity models on top of the dataset, extending coverage as Hercules onboards more retailers.
Business, technical, and governance outcomes.
- Continuous, web-scale FMCG price intelligence in production.
- Cross-store SKU matching that holds the receipt-level lowest-price guarantee.
- Dispatch team always knows where each item is cheapest right now.
- Foundation for promotion detection and price-trend forecasting.
- Python
- Scrapy
- Playwright
- PostgreSQL
- Airflow
- dbt
Scheduled pipelines with data-quality checks, SKU-matching rules versioned in code, and historical snapshots so price claims are reproducible and auditable.
Working on something similar?
Schedule a call. We will tell you honestly whether AI is the right move.
Reference calls available under NDA after the second working session.