AI Design Playbook

A practical guide to design AI features

Get started

Three steps to go from reading to shipping — browse the core chapters, test your instincts, then apply what you've learned to your own product.

1

Learn

Browse the core chapters — jobs, patterns, semantics, visuals, and interaction.

2

Test

Take the AI Design Quiz to check what you know.

3

Apply

Download the Skills and drop them into your own workflow.

Chapters

Focused chapters covering how an AI feature communicates, looks, and behaves — the semantics, visuals, and interaction patterns that make it trustworthy.

Semantics

Content strategy, microcopy, naming, and cues that distinguish AI from human content.

Visuals & Consistency

Component integrity, visual language, iconography, and color patterns for AI.

Interaction & Feedback

The AI interaction loop, loading patterns, and feedback mechanisms that make AI smarter.

Test your AI design knowledge

A three-level, self-graded quiz — thirty questions spanning transparency basics to advanced agentic patterns, drawn from the same checklist used throughout this playbook.

Start the AI Design Quiz

Takes about 10 minutes. No login required.

← Back to home

Semantics

How the feature communicates through language — content strategy, microcopy, naming, and the cues that distinguish AI from human content.

Affordances and control

Make every action explicit: accept, reject, edit, regenerate. Support iteration — let users refine without starting over. Enable graceful exit at any point. Offer granularity — let users control scope ("Rewrite just this paragraph" vs "Rewrite everything"). The goal is partnership, not automation overreach.

Pattern: Refine output

Give users tools to iteratively improve AI output — tone sliders, length controls, style presets, and regenerate buttons. Each refinement should feel like a conversation, not a restart. Users who refine output report 3x higher satisfaction.

Signifiers

Cues that distinguish AI content from user content. They should be present but not loud — a subtle sparkle icon, a light border, or a small label. Keep them consistent across all AI features. Signifiers should degrade after user edits: once a user modifies AI output, the "AI" label fades because it's now their content.

Pattern: AI attribution degradation

AI markers should follow a lifecycle: prominent when first generated → subtle after review → invisible after user edits. This respects user ownership while maintaining transparency during the review phase.

Content transparency

Always distinguish AI-generated from human-curated content. Show drafts as drafts — never present AI output as final without user approval. Be honest about uncertainty and confidence levels. Use visual badges, subtle borders, or labels like "AI-generated draft" to make provenance clear.

Pattern: Human-verified vs. AI-generated

Design clear visual indicators (badges, icons, or color coding) to differentiate AI-generated and human-verified content. Assign confidence levels or accuracy scores. Allow users to filter or sort by verification status.

Microcopy and narration

Set expectations before every action — tell the user what the AI will do before it does it. Narrate the process during loading ("Analyzing your document...", "Generating suggestions..."). Frame results with context ("Based on your last 3 conversations"). Guide feedback with specificity — don't just ask "Was this helpful?" but "Did this match your tone?"

Pattern: Progressive disclosure

Gradually reveal AI capabilities as users become comfortable. Start with simple suggestions, then introduce advanced features. Reduces cognitive overload and builds confidence over time.

Feature naming

Pick a single umbrella brand for AI features (e.g. "[Product] AI") and keep individual feature names in title case: AI Drafts, AI Suggestions, AI Insights. Always prefer verbs over nouns — they communicate action. Be specific about what the AI does. Naming should get sign-off from design and product leadership before shipping.

Pattern: Naming conventions

Consistent naming reduces confusion. Users build mental models around feature names. Changes are expensive — get it right early. Use language that communicates capability without overpromising.

Conversation design

When AI communicates through dialog — chat, voice, or multi-turn interactions — the Conversation Design Institute's framework applies: set expectations early (tell users what AI can and can't do), maintain context across turns (reference previous messages naturally), support mixed-initiative (both user-led and AI-led turns), handle dead ends gracefully (always suggest a next step), and design repair strategies for when misunderstanding occurs.

Pattern: Turn-taking and repair

Good conversations have rhythm. AI should know when to speak, when to listen, and when to ask for clarification. Repair strategies ("Did you mean X or Y?") are essential — they show the AI is paying attention, not just generating. The CDI recommends designing the "unhappy path" first: what happens when the AI misunderstands?

← Back to home

Visuals & Consistency

How the feature looks within your design system — component integrity, visual language, iconography, and color patterns for AI.

Design system integrity

Always check your existing component library first before creating anything new. Extend, don't fork. When a new AI-specific component is needed, document it immediately. Build your AI component system to be extensible from day one: atomic components can absorb confidence indicators, streaming containers, and adaptation signals without a full redesign.

Pattern: Extensible component library

AI features need specialized components (confidence badges, streaming text containers, suggestion cards). Design these as extensions of existing primitives, not new paradigms. This keeps the system cohesive while supporting AI-specific needs.

UX consistency

Similar AI behaviors should look similar across features. A suggestion in one part of the product should feel the same as a suggestion anywhere else. Build a shared interaction vocabulary — if "sparkle" means AI-generated in one place, it means AI-generated everywhere. Standardize the AI flow pattern across all features.

Pattern: Shared interaction vocabulary

Create a consistent visual language for AI interactions: same icons for same actions, same animations for same states, same layouts for same patterns. Users build muscle memory around consistent patterns — breaking consistency breaks trust.

Iconography

Use a small, clear icon set for AI actions. The sparkle icon (✦) is the primary AI signifier — don't overload it for multiple meanings. Always pair icons with text labels; icons alone are ambiguous. Keep AI icons visually lightweight so they don't compete with primary actions.

Pattern: Icon + label pairing

Research shows icon-only interfaces increase error rates by 30%. For AI features where trust and clarity are paramount, always pair icons with labels. The label does the heavy lifting; the icon provides quick visual scanning.

Color and confidence

Use a subtle AI accent color — distinct enough to recognize, muted enough to not overwhelm. Never conflict with semantic colors (red = error, green = success). For confidence indicators, use a calibrated scale: green (high confidence, >85%), amber (medium, 60-85%), gray (low, <60%). Confidence colors should fade after the user accepts the suggestion.

Pattern: Confidence visualization

One of the most important AI UX patterns of 2026. Show confidence as percentages, color-coded borders, or calibrated badges. Use consistent scales across all features. Always explain what factors influence the score. Show uncertainty ranges, not just point estimates.

← Back to home

Interaction & Feedback

How users interact with and respond to the AI — buttons, triggers, states, loading patterns, and the feedback loops that make AI smarter.

Buttons and CTAs

Clear action hierarchy: primary action = safest action. "Review suggestion" beats "Send now". Value-focused copy — tell users what they get, not what to do. Sentence case, no emoji, no exclamation marks. Provide keyboard shortcuts for power users (Cmd+Enter to accept, Esc to dismiss). Always include a "Not now" escape.

Pattern: Safe defaults

The default action should always be the safest one. "Review" not "Send". "Save draft" not "Publish". AI outputs are probabilistic — the UI should reflect this by making cautious actions prominent and irreversible actions require confirmation.

Triggers and inputs

Use explicit triggers for consequential actions — never auto-apply AI changes without consent. Progressive disclosure for complex inputs: start with a simple prompt, offer advanced options behind a toggle. Always show what context the AI can see ("Using your last 5 conversations and knowledge base"). The autonomy dial pattern lets users choose their comfort level from manual to fully automatic.

Pattern: Autonomy dial

Let users control how much the AI does autonomously. From "suggest only" to "auto-apply with review" to "fully automatic." Pre-action: Intent Preview ensures the user defines the plan. In-action: Explainable Rationale maintains transparency. Post-action: Audit Trail provides a safety net.

States and notifications

Five core states: idle, loading, success, error, streaming. Errors must be specific — never "Something went wrong." Say what happened plainly and provide a recovery path. Show low confidence explicitly. For streaming, design for progressive text reveal with a stable layout that doesn't jump. Notify users of background AI completion without interrupting their current task.

Pattern: Streaming and progressive reveal

Streaming reduces perceived wait time by 60% compared to batch responses. Design for stable layout during streaming — reserve space, use skeleton screens, and never let content jump. Show a subtle cursor or typing indicator. Allow users to stop generation mid-stream.

Loading and progress

Duration-based loading treatments: <1s — no loading state. 1-3s — subtle spinner. 3-10s — spinner + contextual messaging. 10-30s — progress bar + abort option + thought process. >30s — move to background + notification. Always show what the AI is doing, not just that it's busy. Skeleton screens reduce perceived wait by 40% vs spinners.

Pattern: Contextual loading narration

Replace generic spinners with stage-specific messaging: "Analyzing your document..." → "Generating suggestions..." → "Finalizing recommendations...". This builds trust and manages expectations. Users who see contextual loading messages rate AI as 25% more trustworthy.

Feedback loops

Every AI interaction should include lightweight feedback — thumbs up/down, star ratings, or "Was this helpful?" Don't ask for feedback every single time; vary the frequency based on context. Collect both explicit feedback (ratings, corrections) and implicit signals (accepted, edited, dismissed). Feed corrections back into the system. Show users how their feedback improved results over time.

Pattern: Continuous improvement loop

The feedback → learning → improvement cycle is what makes AI products get better. Provide consistent ways for users to submit corrections and ratings. Track acceptance rates as implicit feedback. Close the loop by showing users "AI has learned from your feedback" moments.

Explainability

Users don't need to understand how a model works — they need to understand why it did what it did and what they can do about it. Explainability is a design decision, not a technical footnote. Every AI output should answer two questions: Why this? and What can I change?

Explainability as a design layer

Explainability isn't a technical afterthought — it's a design decision. Users don't want a model architecture diagram; they want a plain-language reason for a recommendation. The goal is actionable understanding: enough context to accept, edit, or reject. Design explainability at three levels: inline (tooltips, "Why this?"), on-demand (expandable rationale panels), and deep (full audit trails for power users).

Do

  • Show the top 2-3 factors behind a recommendation
  • Use "Why this?" toggles that expand inline
  • Phrase explanations as cause → effect ("Based on 3 similar campaigns that converted at 12%")
  • Let users drill deeper if they want, but don't force it

Don't

  • Show raw model scores or technical metrics
  • Bury explanations behind multiple clicks
  • Use vague language ("AI-powered recommendation")
  • Assume users will trust output without context
Pattern: Layered explainability

Design three tiers of explanation: (1) Glanceable — a one-line reason shown inline, always visible. (2) On-demand — a "Why this?" panel with 2-3 contributing factors and data points. (3) Deep — a full audit trail with timestamps, data sources, and model version. Most users stay at tier 1; power users appreciate tier 3.

Confidence communication

Every AI output carries uncertainty. The interface must communicate this honestly — not to undermine the AI, but to calibrate user expectations. Overconfident AI erodes trust when it's wrong; appropriately uncertain AI builds trust over time. Match the confidence display to the stakes: a subject line suggestion can be casual; a revenue forecast needs precision.

Do

  • Use calibrated visual scales: green (>85%), amber (60-85%), gray (<60%)
  • Show ranges, not just point estimates ("Expected open rate: 18-24%")
  • Explain what would increase confidence ("More data needed")
  • Fade confidence indicators after user accepts

Don't

  • Hide low confidence — surface it clearly
  • Use confidence colors that clash with semantic colors (red/green)
  • Show raw percentages without context
  • Display confidence on trivial outputs where it adds noise
Pattern: Confidence calibration

Confidence scores should map to real-world accuracy. If the AI says "85% confident," it should be right ~85% of the time. Miscalibrated confidence (saying 90% when accuracy is 60%) is worse than no confidence at all. Test calibration regularly and adjust thresholds per feature. Use text alternatives alongside color ("High confidence" not just a green dot).

Trust

Trust is the currency of AI products. It's earned incrementally — through honest errors, clear data disclosure, reliable escalation paths, and interfaces that never feel inescapable. Every AI feature should make the user feel in control, even when the AI is doing the heavy lifting.

Data transparency and consent

Users need to know what data the AI can see, how it's used, and what control they have. This isn't just compliance — it's a trust multiplier. Show context windows explicitly ("Using your last 5 conversations"), provide clear opt-out paths, and never process data without informed consent. Data transparency should be persistent and accessible, not buried in onboarding.

Do

  • Show what data the AI is using ("Based on your contact list and last 30 days of messages")
  • Provide granular data controls (per-feature, per-data-type)
  • Allow users to delete their AI training data
  • Surface data usage in context, not just in settings

Don't

  • Process data silently without disclosure
  • Use dark patterns to get consent (pre-checked boxes, confusing language)
  • Make opting out harder than opting in
  • Store data longer than needed without explanation
Pattern: Visible context window

Always show what the AI can "see." Display a small, persistent indicator: "AI can access: your flows, contacts, and message history." Let users expand it to see specifics and toggle individual data sources on/off. When the AI accesses new data types, notify the user and ask for permission.

AI attribution and provenance

Users must always know whether content was human-created or AI-generated. Attribution should follow a lifecycle: prominent when first generated, subtle after user review, invisible after substantial edits. This respects user ownership while maintaining honesty during the review phase. For content that will be seen by end-users (subscribers, customers), AI provenance must persist.

Do

  • Label AI-generated content clearly ("AI draft", "Suggested by AI")
  • Fade attribution after user edits >30% of content
  • Maintain provenance in audit logs even after UI labels fade
  • Distinguish between AI-generated, AI-assisted, and human-created

Don't

  • Present AI output as if the user wrote it
  • Remove all attribution the moment a user clicks "Accept"
  • Use the same label for very different levels of AI involvement
  • Let AI attribution compete visually with primary content
Pattern: Attribution degradation

AI markers follow a lifecycle: (1) Prominent — "AI-generated draft" badge when first created. (2) Subtle — small icon after user reviews and edits. (3) Invisible — removed from UI after substantial user modifications, but preserved in metadata. This pattern respects both transparency requirements and user ownership of their content.

Error honesty and recovery

When AI fails, the interface must be honest about it. Vague errors ("Something went wrong") destroy trust; specific errors ("Couldn't generate a subject line because the message body is empty") preserve it. Every error needs a recovery path — what the user can do next. For AI-specific failures (hallucinations, low-quality output), acknowledge the limitation directly.

Do

  • Say what happened in plain language
  • Provide a clear next step ("Try rephrasing" or "Add more context")
  • Acknowledge when the AI doesn't know something
  • Offer a human fallback for critical tasks

Don't

  • Show generic error messages
  • Pretend failures didn't happen (silently retrying without disclosure)
  • Blame the user for AI limitations
  • Let the user hit a dead end with no recovery option
Pattern: Graceful degradation

When AI confidence drops below a threshold, automatically fall back to simpler, more reliable outputs. A text generator might fall back to templates. A classifier might show "Unsure — please review" instead of forcing a category. Design the degradation path explicitly — don't leave it to chance. Users should barely notice the transition.

Human escalation and override

No AI system should be a black box with no exit. Users need the ability to override any AI decision, escalate to a human when the stakes are high, and permanently opt out of specific AI features. The escalation path must be visible and frictionless — not hidden behind settings menus. For agentic workflows, build explicit checkpoints where the user reviews and approves before the AI proceeds.

Do

  • Show "Override" or "Edit manually" on every AI output
  • Provide one-click escalation to human support
  • Remember user overrides and learn from them
  • Build checkpoint gates into multi-step AI workflows

Don't

  • Make the AI feel inescapable
  • Require multiple steps to reach a human
  • Ignore repeated overrides (if users keep correcting the AI, something is wrong)
  • Auto-proceed past checkpoints without explicit user approval
Pattern: Checkpoint gates

For agentic AI that takes multi-step actions, insert mandatory review points before irreversible steps. Show what the AI plans to do, what data it will use, and what the expected outcome is. The user explicitly approves before the AI proceeds. This is the HITL (Human-In-The-Loop) pattern applied at the interaction design level.

Building trust over time

Trust with AI isn't binary — it's earned incrementally. New users need more transparency, more confirmation, and more conservative defaults. As users gain experience and the AI proves reliable, the interface can progressively reduce friction. This is the autonomy ladder: start supervised, earn independence. Track trust signals (acceptance rate, override frequency, feature adoption) to calibrate the experience per user.

Pattern: Progressive autonomy

Map the trust journey: (1) Suggest — AI recommends, user acts. (2) Auto-apply with review — AI acts, user confirms. (3) Fully autonomous — AI acts, user is notified. New features start at stage 1. Users graduate based on demonstrated comfort (high acceptance rates, low override frequency). Always allow users to dial back to a more supervised mode. The autonomy dial should be per-feature, not global.

← Back to home

AI & Agentic Jobs

Every AI feature performs one or more of 7 core jobs. Name the job first — the interaction pattern follows.

Anticipate

See what's coming before the user asks. Uses behavioural and historical patterns to surface insights, flag anomalies, or predict next steps. The highest-uncertainty job — confidence communication is non-negotiable.
Example: Next-best-action suggestions in flows

The design tension: being helpful early vs. being wrong loudly. Err on the side of showing your work.

Do

  • Show confidence levels with every prediction
  • Explain what data informed the forecast
  • Offer alternative scenarios, not just the top pick
  • Update predictions in real time as context changes

Don't

  • Present predictions as facts
  • Hide the reasoning behind projections
  • Rely on thin data without flagging it
  • Auto-act on a prediction without user consent
Pattern: Proactive intelligence

Surface relevant information before the user asks for it, based on behavioural patterns and context. The best anticipation feels like good timing, not surveillance.

Discover

Help users find what they need using natural language. Translates intent into semantically relevant results — not keyword matches. Often the first AI capability users encounter, so first impressions matter.
Example: Semantic search in a message inbox

The design tension: speed of retrieval vs. accuracy of understanding. Always show the AI's interpretation so users can course-correct.

Do

  • Show how the AI interpreted the query
  • Highlight matching context within results
  • Offer one-click refinement suggestions
  • Support conversational follow-up ("show me more like this")

Don't

  • Return only exact keyword matches
  • Show empty states without actionable next steps
  • Require query syntax or special operators
Pattern: Semantic retrieval

Match user intent, not just keywords. Show what the AI understood and let users refine naturally. Zero-result states are a design opportunity, not a dead end.

Generate

Create new content on the user's behalf — replies, drafts, summaries, comments. The most visible AI job and the one users judge hardest. Trust is earned by never pretending the output is final.
Example: AI Drafts, AI Suggestions

The design tension: speed of creation vs. faithfulness to the user's voice. Always present output as a draft, never as done.

Do

  • Frame output as a draft the user controls
  • Offer tone, length, and style controls
  • Show a clear "AI-generated" indicator
  • Allow inline editing before any action

Don't

  • Auto-send without explicit confirmation
  • Override the user's established voice
  • Present AI output as the user's own words
Pattern: Human-AI co-creation

AI drafts, human decides. The output is always a suggestion — never a fait accompli. Users who can edit inline before sending report 3× higher satisfaction.

Connect

Bridge the gap between what the user is doing and what they need to know. Surfaces relevant context — knowledge base articles, past conversations, related data — at the moment it's useful.
Example: Knowledge base references in AI Drafts

The design tension: surfacing enough context to be useful vs. overwhelming with information. Rank ruthlessly, show sources always.

Do

  • Show clear source attribution for every reference
  • Rank by relevance first, recency second
  • Let users expand, verify, and pin useful context
  • Display inline — not buried in popups or modals

Don't

  • Show context without linking to the source
  • Overwhelm with too many references at once
  • Surface stale or irrelevant information
Pattern: Contextual augmentation

Right information, right moment, clear attribution. The best Connect experiences feel like having a teammate who always remembers the relevant detail.

Recommend

Proactively nudge the user toward the next best action. Surface suggestions, highlight opportunities, or call attention to things they might miss. Timing is everything — too early is noise, too late is useless.
Example: AI Insights recommendations

The design tension: being proactive vs. being intrusive. Recommendations should feel like a helpful nudge, never a pop-up ad.

Do

  • Explain the reasoning behind each suggestion
  • Allow one-tap dismissal and snooze
  • Adapt frequency based on user behaviour
  • Use subtle, non-blocking UI (banners, inline cards)

Don't

  • Interrupt the user's current task with a modal
  • Repeat a suggestion the user already dismissed
  • Recommend without explaining why
  • Use dark patterns to push acceptance
Pattern: Ambient intelligence

Anticipate needs without disrupting flow. The best recommendations arrive exactly when the user was about to look for them.

Enhance

Polish what the user already created. Tone-match, tighten copy, fix grammar, improve structure — without overwriting their intent. The goal is refinement, not replacement.
Example: Tone matching, message polish

The design tension: improving quality vs. respecting authorship. Every change should be visible, and the user should always be able to accept or reject each one.

Do

  • Show a clear before/after comparison
  • Allow partial acceptance — change-by-change
  • Preserve the user's original meaning and voice
  • Offer multiple refinement directions (shorter, friendlier, more formal)

Don't

  • Replace content silently without showing the diff
  • Change the meaning or intent of the original
  • Auto-apply refinements without review
  • Impose a voice that doesn't match the user's
Pattern: Collaborative editing

Enhance what exists. Show every change. Let the user stay in control. Users who refine AI output report 3× higher satisfaction than those who accept or reject wholesale.

Organize

Bring structure to chaos. Auto-tag, cluster, categorize, or segment content by finding patterns humans might miss. Turns unstructured data into something actionable.
Example: Auto-tagging conversations by topic

The design tension: automation speed vs. accuracy of classification. Let the AI sort, but always let humans correct.

Do

  • Show why items were grouped together
  • Let users rename, merge, or split clusters
  • Provide easy override controls
  • Display grouping confidence scores

Don't

  • Auto-apply labels without letting users review
  • Create opaque clusters with no explanation
  • Hide the ability to manually re-categorize
Pattern: Auto-organization

Let AI find structure in unstructured data, but always let humans verify and adjust. The best Organize features feel like a smart filing system, not an opinionated robot.

Jobs in practice

Most features combine multiple jobs — that's by design. AI Drafts performs Generate + Connect (drafts a response while pulling knowledge base context). AI Insights does Recommend + Anticipate (suggests what to focus on based on patterns). Start by naming which jobs your feature performs, then design the interaction pattern for each through the 4-stage loop.

← Back to home

Agentic Patterns

Design patterns for autonomous and semi-autonomous AI agents — from the 4-stage interaction loop to delegation, escalation, and multi-step workflows.

Autonomy patterns

Sequential pipeline

A fixed chain where output from step A feeds step B, then step C. Best for known, repeatable processes like onboarding flows, compliance checks, or document processing.

Example: Extract data → Validate → Draft reply → Submit for review

Intent preview + execution

Before taking action, the agent shows a plan: "I'll search your inbox, find relevant threads, then draft a summary." The user approves, modifies, or cancels before anything happens.

Example: User asks "summarize my week" → Agent shows 5-step plan → User approves → Agent executes with progress

Autonomy escalation

Start with AI suggesting, then auto-applying with review, then fully automatic — as user trust builds over time. The autonomy dial adjusts per feature and per user.

Example: Week 1: AI suggests replies → Week 3: AI drafts + asks approval → Month 2: AI sends with undo window

Checkpoint + rollback

For multi-step workflows, save state at each step so users (or the system) can roll back to any point. Never let an agent failure cascade into lost work.

Example: Research → [checkpoint] → Draft → [checkpoint] → Format → [checkpoint] → Send

Human-in-the-loop gates

Require explicit human approval before irreversible actions. The agent pauses, presents what it's about to do, and waits. No timeouts that auto-approve.

Example: Agent drafts email → Shows preview → Waits for "Send" click → Never auto-sends

Readiness framework

Before building an AI feature, walk through these decision gates:

1. Is the task repetitive and time-consuming for users?

AI adds most value for high-volume, repetitive tasks. If the answer is no, reconsider whether AI is the right solution.

2. Is the AI output verifiable by the user?

If users can't verify correctness, AI may cause more harm than good. Proceed with caution.

3. Can the user recover from AI mistakes easily?

Low-recoverability actions need stronger guardrails or human-in-the-loop gates.

4. Does the AI have enough data to be useful?

Don't ship AI that performs worse than the non-AI version. Insufficient data = bad results.

5. Will users trust AI for this specific task?

High-stakes tasks need more transparency and gradual rollout. Lower-stakes can be more ambient.

← Back to home

Skills

Drop these into Claude Code or Cowork to get instant AI design reviews and content checks.

Available skills

Skills are downloadable configuration files that give AI tools specialized knowledge about this design system and guidelines. Install them to get context-aware design feedback during your workflow.

← Back to home

AI Design Quiz

Test your knowledge of AI feature design — transparency, user control, content, trust, and accessibility.

How well do you know AI design?

30 questions across 3 levels, drawn from the AI feature design checklist. Pick a level and see how you score.

Foundations
Intermediate
Advanced
Question 1 of 10
How should AI-generated content be presented to users?
Mixed in with human content to feel seamless
Visually distinguished from human content
Hidden behind a toggle so it doesn't distract
AI output should always be visually distinguishable — a subtle icon, border, or label — so users know what was generated vs. what they wrote.
Question 2 of 10
What actions should users always have on AI output?
Accept or reject only
Accept, reject, and share
Accept, edit, reject, and regenerate
Full control means accept, edit, reject, or regenerate. Users should be able to refine AI output iteratively, not just take-it-or-leave-it.
Question 3 of 10
When should confidence levels be shown to users?
Always, on every AI output
When relevant — especially for ambiguous or uncertain results
Never, it makes the AI look unreliable
Confidence levels help users calibrate trust. Show them when the AI's certainty matters for the decision at hand — not on every trivial output.
Question 4 of 10
What should happen when a user starts an AI action and it takes more than 2–3 seconds?
Show what the AI is doing and provide an abort option
Show a generic spinner until completion
Hide the loading state to avoid anxiety
Contextual narration ("Analyzing your data…") plus an abort button after 2–3 seconds gives users control and reduces uncertainty.
Question 5 of 10
What's the recommended casing for AI feature names?
ALL CAPS — AI DRAFTS
Lowercase — ai drafts
Title case — AI Drafts
Title case reads as a proper feature name, not a shouted label or a raw identifier. Prefer verbs over nouns in the name — they communicate what the feature does.
Question 6 of 10
Should AI be allowed to auto-execute irreversible actions?
Yes, if the AI is confident enough
No — irreversible actions always need explicit user confirmation
Only during onboarding to showcase AI power
No irreversible AI actions without explicit confirmation. The agent should pause, preview, and wait — never auto-approve.
Question 7 of 10
How should error messages in AI features be written?
Specific, with a recovery path the user can take
Generic — "Something went wrong. Try again."
Technical — showing the API error code
Error messages should be specific about what went wrong and offer a clear recovery path. Generic errors leave users stranded.
Question 8 of 10
How should CTAs in AI features be written?
Action-focused — "Run AI"
Technical — "Execute model inference"
Benefit-focused — tell users what they'll get
CTAs should be benefit-focused: "Get suggestions" beats "Run AI". Tell users what value the action delivers.
Question 9 of 10
What should AI never do with probabilistic output?
Show it to users
Present it as certainty
Include confidence scores
AI must never present probabilistic output as certain fact. When the AI isn't sure, it should say so — hedging language, confidence bars, or qualifiers.
Question 10 of 10
What contrast ratio must AI features meet for body text (WCAG AA)?
4.5:1
3:1
2:1
WCAG AA requires 4.5:1 contrast for body text and 3:1 for large text. AI features are no exception.
0
0 of 10 correct
Question 1 of 10
What is "contextual narration" in AI loading states?
A voiceover explaining the AI feature
Dynamic text showing what the AI is currently doing
A tooltip that appears on hover
Contextual narration means the loading state updates to reflect the current step — "Reading your data…", "Generating suggestions…" — instead of a generic spinner.
Question 2 of 10
What is the "autonomy dial" concept?
An AI model parameter that controls creativity
A setting for how often AI provides suggestions
A user control that adjusts how aggressively AI acts on their behalf
The autonomy dial lets users control AI aggressiveness — from "suggest only" to "act on my behalf". It respects different comfort levels with automation.
Question 3 of 10
Why should feedback requests be varied (not asked every interaction)?
Over-asking causes feedback fatigue and lower response quality
It's more cost-effective to ask less often
Users don't have opinions on most interactions
Feedback fatigue is real. If you ask after every AI interaction, users start ignoring the prompts. Vary timing and use implicit signals too.
Question 4 of 10
What should "source attribution" look like in AI output?
A generic "Powered by AI" label
Specific references to the data or information the AI used
A link to the AI model's documentation
Source attribution means showing which specific data, documents, or context the AI accessed to generate its response — not just "AI made this".
Question 5 of 10
Which words should be avoided in AI feature microcopy?
"Smart" and "intelligent"
"Automatic" and "instant"
"Train" and "fine-tune"
Avoid "train" and "fine-tune" — these imply the user is doing ML work. Use terms like "improve", "teach", or "customize" instead.
Question 6 of 10
What should streaming/progressive reveal be used for?
Long AI responses — showing output as it's generated
Short responses to make them feel more dramatic
Loading screens to entertain users
Progressive reveal works for long responses where the user can start reading while the AI continues generating. It reduces perceived wait time.
Question 7 of 10
What does "AI attribution degradation" mean?
AI labels becoming less accurate over time
AI markers fading after the user edits the output — because it's now their content
Reducing the quality of AI output to save costs
Once a user modifies AI output, the "AI" label should fade because it's now their content. Prominent when generated → subtle after review → invisible after edits.
Question 8 of 10
What should confidence indicators include besides color coding?
Animation effects
Sound cues
Text alternatives for accessibility
Confidence indicators must not rely on color alone. Always include text alternatives — labels, percentages, or descriptive text — for users with color vision deficiency.
Question 9 of 10
What is a human escalation path?
A way for users to reach a human when the AI can't help
A process for humans to review AI model performance
An internal workflow for AI engineers
Every AI feature should have a human escalation path — a clear way for users to reach a real person when the AI falls short or the situation is too complex.
Question 10 of 10
What should streaming AI content use for screen reader users?
Auto-focus management
aria-live regions
Sound notifications
Streaming content should use aria-live regions so screen readers announce new content as it appears, without disrupting the user's current focus.
0
0 of 10 correct
Question 1 of 10
In an agentic AI pattern, what should happen before an irreversible multi-step action?
The agent should auto-execute quickly to save time
The agent should ask the user to confirm each micro-step
The agent should pause, preview its full plan, and wait for explicit approval
Human-in-the-loop gates: the agent pauses, presents its intended actions, and waits for approval. No timeouts that auto-approve. The user decides.
Question 2 of 10
What's the best approach when the AI needs to handle both low-stakes and high-stakes tasks?
Treat all tasks the same for consistency
Low-stakes can be ambient; high-stakes need more transparency and gradual rollout
Only enable AI for low-stakes tasks
Risk calibration matters. Auto-suggesting a reply draft (low-stakes) is fine to be ambient. Sending a mass email (high-stakes) needs preview, confirmation, and undo.
Question 3 of 10
When should you NOT ship an AI feature?
When it performs worse than the non-AI version
When it takes more than 3 seconds to respond
When it requires user onboarding
If the AI doesn't beat the baseline (non-AI experience), don't ship it. Insufficient data = bad results = worse experience.
Question 4 of 10
How should bias be handled in AI suggestions and language?
Rely on the model provider to handle it
Add a disclaimer that AI may be biased
Proactively audit for bias in suggestions and language before shipping
Bias auditing is a pre-ship requirement, not an afterthought. Test across demographics, review language patterns, and set up ongoing monitoring.
Question 5 of 10
What is the correct approach to "Why this?" explanations?
Show the raw model weights and parameters
Make them available on-demand so users can understand the AI's reasoning
Only show them to power users in settings
"Why this?" explanations should be accessible to all users on-demand — not forced, not hidden. They build trust by letting users peek behind the curtain when they want to.
Question 6 of 10
What's the relationship between user data usage and trust?
Data usage must be transparent and controllable by the user
Data usage should be hidden to avoid confusing users
Users should accept data usage as part of using AI
Trust requires transparency. Users should see what data the AI accesses, and have controls to limit or revoke that access.
Question 7 of 10
What should the release cadence look like for AI features with uncertain user trust?
Full launch to all users immediately
A/B test with 50% of users
Gradual rollout — start with suggest-only, increase autonomy as trust builds
Gradual rollout lets trust build organically. Start with suggestions, let users opt into more autonomy, and expand based on feedback and metrics.
Question 8 of 10
How should guardrails be designed for AI outputs?
Block all edge cases and show a generic error
Prevent harmful or inappropriate outputs while preserving useful functionality
Let the model handle safety on its own
Guardrails should be precise — preventing harm without over-blocking legitimate use. Combine model-level safety with product-level constraints and human escalation.
Question 9 of 10
For keyboard accessibility, what must be true of all AI suggestions?
They must be fully navigable and actionable via keyboard alone
They need a keyboard shortcut to dismiss them
They should auto-focus when they appear
All AI suggestions must be keyboard-accessible — navigable with Tab/arrows, actionable with Enter/Space. Mouse-only AI features exclude keyboard and screen reader users.
Question 10 of 10
What makes a task a good candidate for AI automation?
It's complex and requires creative judgment
It's new and users haven't learned it yet
It's repetitive, time-consuming, and the output is verifiable by the user
AI adds the most value for high-volume, repetitive tasks where users can verify the output. If users can't check correctness, AI may cause more harm than good.
0
0 of 10 correct