#77 – The New UX – Designing for AI Agents

For the entire history of software, UX design has been about human mechanics and psychology: Where do we put the button? How do we make the navigation intuitive? But a massive shift is underway. What happens to User Experience when the user is no longer the one clicking the buttons, but instead is managing an autonomous AI agent doing the clicking for them?

In this episode, we explore the wild new frontier of „Agentic UX.“ We dive into the profound shift from designing for interaction to designing for trust and delegation. When an AI agent goes off to book your flights, negotiate a contract, or refactor your codebase in the background, traditional UI components are useless. Instead, designers must figure out how to visualize the invisible, keep the human „on the loop“ without overwhelming them, and seamlessly handle the moments when the AI gets confused and needs to ask for human permission.

In this episode, we unpack:

• The „Trust by Design“ framework: How to build interfaces that transparently show how and why an AI agent is making decisions in the background.

• The Visibility Problem: Designing real-time „progress states“ for complex, multi-hour autonomous tasks so the user doesn’t panic and hit cancel.

• Handoffs and Escalations: How to elegantly design the exact moment an AI agent hits a roadblock and needs to hand the steering wheel back to a human.

• Machine-to-Machine UX: Why the best user experience of the future might just be an incredibly clean, perfectly structured API meant only for other AI agents to read.

The future of design isn’t about making software easier to use; it’s about making software easier to manage. Tune in to find out how to design for the era of autonomous digital workers.