Case Study: Mangoes.ai achieves 4x faster, deterministic healthcare voice responses with Redis LangCache

Mangoes.ai delivers 4x faster healthcare answers with Redis

Mangoes.ai, a provider of a voice-first AI assistant for behavioral healthcare, faced the challenge of delivering high-quality, consistent, and fast answers to caregivers without incurring exorbitant LLM costs. They needed deterministic responses that caregivers could trust and a solution that would scale affordably in production.

By implementing Redis LangCache, a managed semantic cache from Redis, Mangoes.ai routes queries to check for semantically similar matches before calling an LLM. This solution achieved a 70% cache hit rate, which resulted in a 70% savings on LLM spend. It also delivered responses four times faster, significantly improving the real-time conversational experience for its users.