For decades, app architecture and API design rested on a single assumption: a human user is on the other side, clicking, tapping, and scrolling through pre-built views. That assumption is no longer valid. With AI agents and automated workflows now accounting for over 50% of web traffic, traditional User Experience (UX) is evolving into Agentic Experience (AX) an architecture built for humans who delegate intent and AI agents that execute actions in milliseconds.
Moving from UX to AX does not mean abandoning the visual interface for text-only chat-boxes, it requires new open-source protocols that redefine how interfaces are discovered, negotiated, and rendered.
In this session, we will look at some real world examples while diving deep into the emerging open protocol stack powering the agentic experience:
Existing Agent Contract: How existing accessibility trees provide the foundational semantic layer that browser agents rely on to parse state and intent.
WebMCP: How the emerging browser specification and hooks let client-side applications register deterministic tools and state directly with in-browser and embedded agents.
MCP Apps: How the Model Context Protocol extends beyond data so servers can ship interactive, sandboxed UI surfaces directly into AI chat.
A2UI (Agent-to-UI) Protocol: How declarative JSON schemas replace fragile LLM-generated HTML, enabling agents to stream secure, native, design-system-aligned components across web and mobile surfaces.