Using Vision LLMs to Understand 500,000 Architectural Drawings

Real estate portfolios are filled with millions of unstructured documents — from building plans to inspection reports. For decades, this data has been trapped in PDFs and CAD files, making it very challenging to extract insights at scale.

In this session, we explore how Vision Large Language Models (LLMs) are unlocking this hidden potential. By combining visual understanding with textual context, AI can now interpret architectural drawings, identify floor layouts, connect spatial data with leases, energy systems, and maintenance workflows.

Findable shows how they’ve used AI to process over 500,000 architectural drawings, turning unstructured documentation into a structured, searchable foundation for better building management, sustainability insights, and smarter design decisions.

‍