Snowflake LLM support Hey folks, I’m currently working on building a scalable, LLM-driven reporting system within Snowflake using Cortex Analysts and a Streamlit application. The setup includes \~14 agents (from data gathering and transformation to visualization and insight narration), each responsible for a specific task in the pipeline. At the moment, I’m facing a few challenges: The generated report seems to be partially hardcoded (\~50%) and partially LLM-driven, and I want to make it fully dynamic and scalable. Additionally, CoCo seems to be modifying some files, which is reducing my confidence in the transparency of the pipeline. I need to make sure the report is generated completely with agents and LLM response and needed your support if you can help in this & is accurate as per the dataset to reduce hardcoded logic in snowflake .…