Over 200 project managers, engineers, and staff in FCS navigate Cornell’s increasingly complex building systems. Institutional knowledge stored across 60 PM Checklists, 47 Guidance Documents, 26 Templates, and 1,000+ pages of Design Standards across 24 trades. The challenge is making this vast, multimodal knowledge base accessible. When a user asks a question, the AI tool retrieves relevant sections and then provides a detailed answer along with links to relevant documents from the knowledge base. For image-heavy documents, the tool generates textual descriptions to bridge image-based and text-based retrieval. The result is prompt and comprehensive answers; enhanced accessibility to knowledge; substantial time efficiencies. Over the longer term, this effort supports faster onboarding, prevention of costly construction mistakes, and improved compliance with design standards.