Most architecture documentation lives in slide decks and wikis — formats that humans struggle to act on and LLMs can't reason over reliably. This talk introduces CoDL (Constraints Description Language) and CaDL (Capabilities Description Language) as lightweight, structured notations for expressing architecture in a form that both governance processes and AI tooling can consume. Drawing on the BTABoK's Architecture Description competency — which emphasises producing structured, stakeholder-relevant, and traceable representations of systems — the session shows how formalising architectural intent into machine-readable schemas unlocks new possibilities: automated compliance checks, LLM-driven design critiques, and governance workflows that run without manual chasing.
Attendees leave with a working mental model of what these languages look like, where they slot into everyday architecture work, and why getting the notation right is the prerequisite for everything else in the AI-assisted architecture stack.