All UI UX practice
UI UX practice · AI UX

AI UX design practice: products people can trust

AI features fail in new ways: they act without asking, sound sure when they are wrong, or leave people stuck. Each challenge below is one of those moments.

Start the first challenge

Designing for AI is less about the screen and more about trust. When should the AI act on its own, and when should it ask first? How does it show that it might be wrong? What happens when it fails?

These challenges put you in front of real-looking AI products, from a travel assistant that books the wrong train to a research tool that invents a source. You choose what the product should do, then see the reasoning and the principle behind the best answer.

The challenges

Read the problem here, then solve it in UX Quest. Beginner challenges are free with an account; the others are part of Pro.

Beginner · AI Trust & InteractionFree

AI Travel Assistant — Wrong Train Booking

An assistant booked a non-refundable 06:12 train after an ambiguous request for a 'flexible morning trip'.

Solve it free Calibrated Trust
Beginner · AI Trust & InteractionFree

Smart Calendar — Unexpected Auto-Reschedule

A calendar agent moves an important meeting based on incomplete context.

Solve it free Calibrated Trust
Intermediate · States Pro

AI Assistant — the blank chat box no one starts

An AI assistant opens to an empty input with a blinking cursor and nothing else. First-time users stare, unsure what to even ask, and leave within a minute.

Solve it Aesthetic-Usability Effect
Intermediate · States Pro

AI Copilot — a popup on every keystroke

An AI coding copilot fires a suggestion modal, a 'tip', and an upsell the moment you start typing. It breaks flow so often that developers reflexively dismiss everything — including the suggestions that would have helped.

Solve it Interruption Cost
Intermediate · AI Trust & Interaction Pro

Email Assistant — One-Tap Reply, Wrong Tone

A one-tap smart reply sends instantly, in a tone that doesn't fit the thread.

Solve it Calibrated Trust
Intermediate · AI Trust & Interaction Pro

Health Assistant — Unsafe Advice

A health assistant gives overly confident advice without explaining limits or directing the user to professional help.

Solve it Calibrated Trust
Advanced · AI Trust & Interaction Pro

Research Assistant — A Confident, Fake Citation

An AI answer includes a specific-looking source that doesn't exist.

Solve it Calibrated Trust
Advanced · AI Trust & Interaction Pro

Support Bot — No Way to Reach a Human

A support bot loops the user through canned answers with no path to a human.

Solve it Calibrated Trust
Advanced · AI x UX Pro

AI Summaries — the ones that quietly hallucinate

An AI feature generates summaries, and users occasionally get a confident, wrong one with no way to tell — design the loop that catches it before they trust it.

Solve it Calibrated Trust

How to approach these challenges

  1. Ask what happens if the AI is wrong here, and how hard that is to undo.
  2. Let the AI act alone only on things that are easy to undo; confirm the rest first.
  3. Show how sure the AI is, and where its answer comes from.
  4. Design the empty, loading and error states, not just the happy path.
  5. Always leave a way out: undo, edit, or reach a human.

AI UX: common questions

  • AI UX design is designing how people use AI features: when the AI acts or asks, how it shows uncertainty, how people correct it and what happens when it fails.

  • Before anything that is hard to undo or affects other people, such as booking, paying, sending or deleting. Small, reversible actions can happen on their own with an easy undo.

  • Assume they will happen. Show confidence and sources, make outputs easy to check and edit, and always offer undo or a way to reach a person.

  • The travel and calendar challenges are free Beginner challenges. The others are Intermediate or Advanced, part of UX Quest Pro.

The UX laws behind these challenges

  • Calibrated TrustShow users when to trust, verify, and override AI output.
  • Aesthetic-Usability EffectUsers perceive aesthetic designs as more usable, but aesthetics alone won't carry a confused user.
  • Interruption CostUnsolicited modals break flow and erode trust. Reserve them for irreversible decisions.
  • AI product design: 8 principles for AI features people trustGuide
  • Heuristic evaluation: how to run one in an afternoonGuide
UX QuestUX Laws and PrinciplesUI UX PracticeUX Design Challenge GeneratorNotion TemplatesBlogPricingAboutContactPrivacy PolicyTerms of ServiceRefund PolicySitemapChangelogPopular UX laws:60-30-10 RuleDesign every stateStatistical PowerFunnel TriageAccessible contrastDensity Matches ContextHick's LawFitts's LawAll UX laws