User Study

Generative UI Report 2025

A study with 145 people from 5 countries exploring how
Generative UI is redefining the AI native experiences.

Introduction

Clicks → Conversations

AI apps are redefining digital life with flexible, prompt-driven workflows that adapt to replace multiple apps.

Back to the Terminal

AI is powerful, yet we’ve gone from rich graphical experiences to text-based chatbots. The modern-day terminals.

What is Generative UI?

Generative UI augments LLMs to respond with live GUI components that adapt to each query and context.

This report explores the before and after of users' experiences when Generative UI is introduced in general purpose chatbots across a range of queries.

The Problem

Shortcomings of today's interface of AI chat tools and how users cope.

The Shift

Drivers of user engagement, preferences and efficiency in AI chat tools.

The Impact

How Generative UI reshapes perceptions and behaviors in AI use.

Insights

1

Text-only responses are overwhelming and hard to follow

2

Structured outputs make information more usable

3

Users leave AI apps when outputs lack structure, credibility, and supportive visuals.

4

When AI generated responses are presented well, users are more engaged

5

When follow-through requires excess effort, task completion suffers

6

Generative UI reduces the effort of using AI effectively, enhancing task-efficiency

Information Overload

Text-only responses are overwhelming and hard to follow

When does more information start to feel like too much? Users describe feeling flooded by detail, skimming for relevance, and risking missed meaning while just trying to make sense of it.
Information Design

Structured outputs make information more usable

What if AI responses could feel lighter to read without losing depth? Small shifts in how information is shown might change how users focus, explore, and decide what matters.
Expectation Gap

Users leave AI apps when outputs lack structure, credibility, and supportive visuals.

Users can tweak what they ask, but not how outputs appear. When text feels heavy or numbers blur together, they turn elsewhere for clarity. Maybe the problem isn’t the content itself, but the limits of how it’s presented.
Engagement

When AI generated responses are presented well, users are more engaged

What happens when AI responses become something you interact with, look at, not just read? As shift occurs, engagement starts to feel different.
Interaction Cost

When follow-through requires excess effort, task completion suffers

Is using AI agents supposed to feel this hard? Many users treat prompt-writing like a skill test rather than a dialogue, shaping their words carefully and hoping the output lands just right.
Efficient Workflows

Generative UI reduces the effort of using AI effectively, enhancing task-efficiency

When information is easy to absorb, decision-support is efficient, users can focus better and complete tasks more smoothly.

What's Next?

The Road Ahead

Generative UI works best when it balances novelty and familiarity, aligning content, intent, and interface for a seamless experience.

Changing Roles of Teams

Teams shift from execution to strategy: designers guide AI layouts, developers ensure high-performing, aligned outputs.