Case Study: eBay achieves personalized, context-aware conversational shopping on Google Assistant with Neo4j

Neo4j Helps Make Conversational Commerce Possible with eBay ShopBot

eBay, a global ecommerce leader, faced a common problem: standard search boxes and recommendation engines often lose the conversational context behind a shopper’s request (e.g., “camping in Lake Tahoe next week, we need a tent”), forcing users to sift results manually. eBay’s goal was to build a real-time recommendation engine that understands and learns from contextual language to surface the right products quickly.

To solve this, eBay built the eBay App for Google Assistant using a Neo4j knowledge graph combined with in-house natural language understanding, machine learning and predictive models. The graph stores the product catalog and shopper interactions so the bot can ask clarifying questions, traverse relationships, and check inventory in real time—delivering personalized recommendations, millisecond responses at internet scale, resilient clustering, and smooth handling of spelling/grammar; the app runs in cloud Docker containers on Google Assistant with plans to expand to other platforms.

eBay

RJ Pittman

SVP, Chief Product Officer