NEWS · COMPANIES · #1915
Orca Security builds 'Discovery AI' natural-language Sonar query agent on Amazon Bedrock
Orca Security developed Discovery AI, a multi-stage natural-language query agent that translates plain-English questions into valid Sonar queries and runs on Amazon Bedrock. Using a RAG + schema-aware pipeline and model routing via Bedrock, Orca reports reducing query costs by 90%, input tokens by 93%, and latency by 75% versus their initial baseline.
KEY POINTS
- Orca Security developed Discovery AI, a multi-stage natural-language query agent that translates plain-English questions into valid Sonar queries and runs on Amazon Bedrock.
- Using a RAG + schema-aware pipeline and model routing via Bedrock, Orca reports reducing query costs by 90%, input tokens by 93%, and latency by 75% versus their initial baseline.
- This shows a practical enterprise LLM deployment that uses model routing and retrieval to make natural-language security queries fast, cheaper, and more accurate for complex cloud schemas.
WHY IT MATTERS
This shows a practical enterprise LLM deployment that uses model routing and retrieval to make natural-language security queries fast, cheaper, and more accurate for complex cloud schemas.