Chat, voice, and agentic AI are not refinements to existing UX. They are replacements for its first principles. Seven years of AI product design, agentic retail work at Walmart, and a US patent on intent-based recommendation interfaces inform this argument. The core claim: the screen is no longer the primary surface, and the thirty-year vocabulary of menus, modals, and hierarchies is splintering into three distinct paradigms that barely share a grammar.
The chat section alone is worth the full read, not for the conclusion but for the framework. The author synthesizes Erika Hall's 2018 argument from Conversational Design with their own patent on multi-level intent classification to land on a precise rule: structured GUI for known intent, conversational AI for ambiguous intent. That is not a UX opinion. It is a design decision tree. The products that got this right, Notion, Linear, GitHub Copilot, are hybrids. The ones that failed treated chat as a universal interface layer. The author also names the Generative UI pattern, cited from a 2025 arXiv paper, where the interface assembles itself dynamically around user intent, shifting the designer's job from authoring states to authoring the system that generates them.
Voice and agentic AI get equal treatment in sections the excerpt only begins to reveal. Apple's Siri AI announcement in June 2026 is flagged as a structural signal, not a product launch. The deeper question the full piece pursues is what design craft even means when the user stops interacting entirely because the system already acted. That is not a rhetorical question. It is the next decade of the profession.
[READ ORIGINAL →]