Self-Describing Buildings: The Open Data Foundation for Trustworthy HVAC AI 

Future AI-enabled building operations will depend on buildings that can describe themselves. This talk presents Project Haystack as an open common language for building data and Haystack 5, enabled by Xeto, as a foundation for consistent, validatable, machine-usable models. Using HVAC examples rather than code, The session will show how structured data can support trustworthy AI, better fault triage, portable analytics, and future-ready operations while keeping HVAC expertise at the centre