Member-only story

User ExperienceAIUXUIResearch Designing for AI deep research: lessons from Google, Perplexity and Microsoft

Mary Borysova

8 min read9 hours ago

We see AI interfaces every day, and many of them already follow the basic best practices of designing for AI. But what happens when the task is more complex: when generating an answer takes minutes rather than seconds, and the subject of research has many sides to it?

Traditional AI conversations follow a straight line. You ask something, get an answer, then ask a follow-up based on that response. This works fine for simple queries, but breaks down when you’re exploring a complex topic with many interconnected aspects.

In this article I look at interfaces that go beyond the straight line. They answer questions users have when working with AI on complex tasks: Is anything happening? Can I make sense of it? Can I trust this? What do I do next?…