AI assistants are dissolving the boundary between applications, and the design problem is no longer about individual interfaces. The article, published in UX Collective under the title 'Context is the new interface,' argues that human intention rarely maps to a single app, and that Apple's App Intents framework is an early structural answer to this: developers define entities, parameters, and actions so Siri and Apple Intelligence can discover what an app does and invoke its capabilities from outside it. The practical documentation on App Intents, app entity schemas, and onscreen context interpretation is cited directly and is worth reading for anyone building in that ecosystem.
The piece makes a specific and useful conceptual move: it reframes context not as background information but as a design material that must be actively structured. This means deciding what gets remembered across a session, what gets shared between products, what requires user confirmation before the system acts on it, and how incomplete requests should be handled. The restaurant recommendation scenario is used throughout as a concrete thread, and the section on memory is particularly sharp: remembering that a friend recommended a place six months ago is not the same as knowing whether that information is still relevant, and the system should surface its reasoning before acting.
The third argument is about governance, and it is the one most designers will not have thought through yet. When an assistant spans multiple products, questions about which calendar is authoritative, whether an invitation sends automatically, and how errors get corrected do not belong to any single product team. The article frames these as UX governance problems: shared rules that define what the system is allowed to do, when it must ask permission, and how it recovers from mistakes. That framing alone is worth the full read.
[READ ORIGINAL →]