Case Study: Korea Press Foundation improves news search accuracy with Upstage's RAG solution

Korea Press Foundation - Customer Case Study

The Korea Press Foundation (KPF) faced significant challenges with its news search platform, BIG KINDS. Its traditional keyword-based system was ineffective for natural language queries, and the use of LLMs often led to inaccurate, hallucinated answers. To provide reliable, real-time information from its vast database of over 82 million documents, KPF partnered with vendor Upstage for a modern AI solution.

Upstage developed a cutting-edge Retrieval-Augmented Generation (RAG) system for the BIG KINDS AI service. This solution featured a two-stage retrieval process and specialized prompt engineering to deliver precise, context-aware answers. The project was a major success, achieving a quality score of 86 out of 80 and a customer satisfaction rating of 92.2 from the Korea Press Foundation, while also driving an increase in article search traffic.