We’ve spent decades teaching users how to navigate applications through menus, tabs, buttons, and forms. But what if your users didn’t need to navigate at all?
Agentic UI is flipping that model and is bringing about the most exciting shift. Instead of forcing users through fixed screens we build manually, Agentic UI powered applications can interpret user intent and assemble the right interface in real time.
In this talk, we’ll explore how A2UI, Google’s open Agent-to-UI protocol for streaming declarative interfaces, is reshaping application design for mobile and web. You’ll see how natural language understanding, context awareness, tool use, and multimodal input of AI models combine to generate adaptive, platform-native experiences at runtime, to render, update, and personalize those interfaces in real time. We’ll also cover the practical patterns that make this usable in production, including catalogs, surfaces, transport, safety boundaries, fallbacks, and user interaction loops.
By the end of this session, attendees will have a clear blueprint for building rich, intelligent applications that feel less like a maze of screens and more like software that can materialize the right UI when it is needed.
