This talk will discuss how Gen-AI technology is gaining steam in regulated industries across the life sciences domain. This technology presents challenges to utilization in these environments because the LLMs alone often produce erroneous information that cannot be trusted. For example, this talk will focus on creating a Clinical Trial Protocol Design Document, which can be thousands of pages in length and contains significant amounts of scientific information required to design, execute & document a clinical trial. Such documents are produced in accordance with regulatory guidlines from governing agencies (FDA, etc.) so thay require accurate links to a wide range of data sources, complete explainability of content/answers produced, and data security around IP, all while disallowing false or unwarranted claims. We will provide specific insights into an aacelerator we've built called Harmoni, which combines data virtualization, semantic technologies (KGs), traditional analytics/AI, & Gen-AI together to provide clients in the life sciences to utilize Gen-AI tech with confidence. We will provide concrete examples of system improvements using a Graph-RAG approach (KG + LLM) vs RAG by itself.