Case Study: Cresta achieves real-time, low-latency AI contact center scale with Fireworks AI

How Cresta drives millions of real-time, AI-powered contact center interactions with Fireworks

Cresta, a company transforming contact centers with AI, sought a scalable infrastructure solution to power its real-time applications. Their challenge involved handling millions of customer conversations with high availability, minimal latency, and the ability to efficiently customize models for each client without prohibitive costs. They turned to the vendor Fireworks to meet these stringent production requirements.

Fireworks provided the infrastructure for Cresta's Knowledge Assist product, utilizing a single base model with multiple LoRA adapters to enable cost-effective, domain-specific customizations. This solution delivered a 100x cost reduction per inference compared to alternatives, allowing Cresta to scale rapidly. The partnership resulted in Cresta's models outperforming GPT-4 in key tasks while providing the low-latency, reliable performance critical for their enterprise customers.

Cresta

Chuan Wang

Technical Lead Manager