Case Study: Raymond James achieves fast, scalable hybrid search for RAG with Redis

Raymond James boosts AI answer coverage with Redis

Raymond James, a top financial services firm, needed to move from a traditional keyword-based search to a trusted AI-powered platform for its employees. Their initial RAG pipeline using vector search had limitations, as it sometimes surfaced unreliable documents and lacked the precision of keyword matching for financial terminology. They required a fast, scalable hybrid search solution that could combine both methods while meeting strict compliance standards.

By implementing the Redis vector database and its hybrid search capabilities, Raymond James built a new retrieval engine for their platform. The solution enabled fast, multi-vector searches and allowed for custom ranking logic. The results were significantly improved, with user feedback becoming noticeably more positive. Key metrics like question coverage and user engagement improved, and when the system was temporarily reverted, users immediately noticed the degradation in performance, validating the success of the Redis-powered solution.

Raymond James