Reduce RAG costs on Amazon Bedrock with query-aware compression

AWS AI blog details query-aware compression for RAG workloads on Amazon Bedrock, reducing the input tokens sent to foundation models on every call — a meaningful part of RAG costs at scale.