AI Native Product Design: From Interface Interaction to Agent Experience

As AI evolves from a standalone feature into an "intelligent collaborator" within products, product design is undergoing a significant paradigm shift. In the past, digital product design focused on helping users understand interfaces and complete tasks. In AI‑native products, users may simply express a goal – and the AI understands intent, reasons, invokes tools, offers recommendations, and even performs a series of tasks on their behalf.

Designers must rethink: when the traditional “click‑page‑action” model is no longer the only interaction pattern, how should we actually design the relationship between humans and AI?

This session will center on AI Native Product Design, drawing on my practical experience at Amazon, PayPal, Remitly, and other tech companies, to share my thinking and practices for designing next‑generation intelligent product experiences. The framework covers five parts:

1. Paradigm Shift: The essential differences between GUI‑based products and AI‑native products
We compare traditional linear workflows with the dynamic, non‑linear experience of Agentic AI across four dimensions: Goal, Intent, Interaction, and Outcome.
When users provide only a vague goal and the system must autonomously understand, plan, and execute – the design focus shifts from “page design” to “system behavior design.”

2. Case Study 1: Enterprise AI Agent – shortening multi‑hour workflows to minutes
This case involves redesigning complex enterprise workflows, using AI to help users process information, analyze data, and support decisions. Key highlights:
- How to define the Agent’s role and boundaries
- How to design Human‑in‑the‑Loop mechanisms
- How to balance automation efficiency with user control

3. Case Study 2: Global payments platform merchant product – from chatbot to intelligent collaborator
AI no longer just answers questions – it understands merchant goals, synthesizes information, makes recommendations, and helps users take next steps. Key insights:
- When AI participates in user decisions, how should traditional Dashboards, Navigation, and Workflows be redesigned?
- How can AI and traditional UI work together, rather than simply “replacing UI with AI”?

4. A reusable AI Native Product Design Framework
Built around six core questions:
- Goal – What does the user truly want to accomplish, not what do they want to click?
- Intent – How do we enable AI to understand user intent and handle uncertainty?
- Agency – What does AI do autonomously, and which decisions must be returned to the user?
- Interaction – How do we design entirely new human‑AI interactions beyond the interface?
- Trust & Control – How do we let users understand AI behavior, intervene, modify, or undo at any time?
- Outcome – How do we shift from “task completion” to measuring real outcomes for the user?

5. Role Redefinition: From Interface Designer to Intelligence Experience Architect
Designers are no longer responsible only for interface, flow, and visuals – they also need to define AI behavior, capability boundaries, decision mechanisms, and human‑AI collaboration. This is the core capability upgrade required of product designers in the AI era.

When AI starts “doing things for users” rather than “helping users operate,” interface design takes a back seat, and system behavior design becomes central. This talk offers no abstract concepts – instead, it delivers two real cases and a reusable framework to help you make the leap as a designer in the AI age.

1、Understand the paradigm – the fundamental differences in design logic between AI‑native and traditional digital products
2、Take away a framework – a directly applicable methodology for your own AI product design practice
3、Redefine your role – a new perspective on the designer’s position and capability upgrade path in the AI era