With a reliable foundation in place, the team turned its attention to its data and commerce stack. Using BigQuery as its central business intelligence platform, The Nile Group powers dashboards for everything from supply chain and procurement to marketplace management and data enhancement, cutting the time to process massive supplier stock feeds from a full day to just 10 seconds by intelligently processing only the daily changes.

For its dynamic pricing system, the company uses Bigtable to ingest and process multiple, simultaneous data streams in real time. This includes constant updates on supplier stock levels, price changes from publishers who use dynamic pricing, and competitor data. In an approach the team likens to high-speed trading, the system processes up to 5,000 of these updates per second. Before this, processing delays meant some products were sold at a loss, while the real-time speed helps protect margins and reduce order cancellations.

Using the Gemini models on Vertex AI, the company began automating tasks that were previously impossible. For example, with thousands of new books released every week, the team can now automatically reformat inconsistent, publisher-provided descriptions, making them clean and readable for customers. The team is also encoding the deep knowledge of its book buyers by having them create nuanced, descriptive tags like "Nordic Crime." Gemini then analyzes and applies these tags to millions of relevant books, allowing customers to discover titles in a specific, serendipitous way that mirrors an expert, in-person recommendation.