This book is licensed for the personal use of the buyer. Please do not share or resell the file. Short quotations with attribution are welcome.
The information in this book is provided for education. It is not legal, financial or other professional advice. Product and company names are used only to describe them and belong to their owners.
This is a free sample: the introduction and first chapter.
IntroductionIntroduction
An AI feature succeeds or fails in its interface. The same model can delight users in one design and frustrate them in another. Users need to understand what the feature can do, judge when to trust it, recover easily when it is wrong, and stay in control of outcomes that matter to them. None of this happens by accident.
This book shows designers how to create experiences for systems that can be wrong. It covers onboarding and expectation setting, input design and prompting interfaces, presenting outputs with sources and uncertainty, error handling and correction, feedback, conversational and agentic interfaces, accessibility, and research methods for AI products. Each chapter offers patterns, anti-patterns and checklists.
The book is tool-agnostic and written for UX and product designers, researchers, writers and product managers.
How the book is organised
Chapters 1 to 9: principles and mental models for AI experiences.
Chapters 10 to 20: onboarding, inputs and prompting interfaces.
Chapters 21 to 33: presenting outputs, sources, uncertainty and errors.
Chapters 34 to 46: conversation, agents, personalisation and feedback.
Chapters 47 to 56: accessibility, inclusion, dark patterns, research, measurement, design systems, writing and collaboration.
Chapters 57 to 61: a worked summary feature, expertise levels, voice, children and vulnerable users, and cross-cultural design.
Chapters 62 to 68: case studies and checklists.
Chapter 1Why AI Changes Design
Traditional interfaces present deterministic functions: press a button, get a predictable result. AI interfaces present probabilistic collaborators: results vary, may be wrong and improve with input. Designers must design not only for success paths but for uncertainty, error and learning.
You have reached the end of the free sample. The full book continues with 67 more chapters, plus the appendices.