Case Study: Relevance AI achieves 99.5% faster vector search with Redis

Relevance AI gets 99.5% faster with
Redis-powered vector search

Relevance AI, which builds AI agents for customer support, sales outreach, and marketing automation, was struggling as demand grew because its in-house vector database could not deliver the speed and reliability needed for its RAG system. The slow vector searches created bottlenecks, increased internal maintenance work, and limited the performance of customer-facing AI agents.

Relevance AI implemented Redis for high-speed, sub-millisecond vector search and caching, replacing its slower internal setup and outperforming other tested options. With Redis, search times dropped by 99.5% from 2 seconds to 10 milliseconds, improving real-time AI responses, enabling automation at scale, and helping Relevance AI fully automate parts of its own sales and support workflows.

Relevance

Jacky Koh

Co-Founder and CEO