Architecture is about the things that are expensive to get wrong, and about getting them right. Decisions have to be made. This way or that way? Now or later? Do we even need a decision or can we reduce the imapct to the point where 'getting it wrong' is not expensive? Making these decisions is not enough though. Too often well intentioned arhcitectural decisions are made, but never implemented. Maybe they are impractical, maybe they are not understood or simply not accepted by the teams.
This talk will show how to decide when an architectural decision is needed and how to make and document those decisions while ensuring they are practical, actionable, get the right people on board and are understandable to people and to AI agents. The techniques used include collaborative modelling, influence strategies, architecture decision records and agent skills.