ai-interaction-patterns
Skill0 starsChanged 7 months agoApache-2.0View on GitHub
AI-specific interaction design patterns covering wayfinding, prompt UX, human-in-the-loop controls, trust & transparency, AI identity, and context management. Based on Shape of AI (shapeof.ai). Use when asking about 'AI UX', 'AI interaction', 'prompt UX', 'AI trust', 'AI disclosure', 'AI avatar', 'AI personality', 'AI memory UX', 'action plan UX', 'stream of thought', 'AI citations', 'AI controls', 'AI wayfinding', 'AI suggestions', 'gallery pattern', 'follow-up pattern', 'draft mode', 'AI variations', 'AI consent', 'AI caveat', 'human-in-the-loop', 'AI transparency', 'AI state', 'prompt design', 'AI onboarding', or 'generative UI'.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- alexandre-darmon/ai-pm-skillstheir own version, first seen Nov 15, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- alexandre-darmon/ai-pm-skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository37 other files it gives its agents.
- .claude/rules/agents.mdAGENTS.md
- .claude/guides/claude-code/CLAUDE.mdCLAUDE.md
- CLAUDE.mdCLAUDE.md
- core-sdkSkill
- dataflowSkill
- nexusSkill
- kaizenSkill
- kailash-mcpSkill
- cheatsheetsSkill
- development-guidesSkill
- nodes-referenceSkill
- workflow-patternsSkill
- deployment-gitSkill
- frontend-integrationSkill
- testing-strategiesSkill
- architecture-decisionsSkill
- code-templatesSkill
- enterprise-infrastructureSkill
- validation-patternsSkill
- gold-standardsSkill
- security-patternsSkill
- flutter-patternsSkill
- interactive-widgetsSkill
- enterprise-ai-uxSkill
- conversation-uxSkill
- uiux-design-principlesSkill
- value-auditSkill
- eatp-referenceSkill
- coc-referenceSkill
- 29-pactSkill
- claude-code-patternsSkill
- error-troubleshootingSkill
- kailash-mlSkill
- kailash-alignSkill
- co-referenceSkill
- managementSkill
- projectSkill
The file
Files
AI Interaction Patterns
AI-specific UX patterns for designing interfaces where users interact with AI models. Covers the full interaction lifecycle: from first prompt to output verification, memory persistence, and trust building.
Source: Based on Shape of AI pattern library (CC-BY-NC-SA) by Emily Campbell.
How This Differs from Other UI/UX Skills
| Skill | Focus |
| 23-uiux-design-principles | Layout, hierarchy, responsive design (framework-agnostic) |
| 21-enterprise-ai-ux | Enterprise context: challenge taxonomy, professional palettes, RBAC |
| 22-conversation-ux | Thread management, branching data model, context switching |
| 20-interactive-widgets | Widget protocols, rendering pipeline, state management |
| 25-ai-interaction-patterns (this) | AI-SPECIFIC interaction logic: how users prompt, control, trust, and relate to AI |
Reference Documentation
- ai-interaction-patterns - Full reference for all 60+ patterns
Quick Pattern Selection
By User Problem
| User Says/Feels | Apply Pattern |
| "I don't know what to ask" | Gallery, Suggestions, Templates |
| "AI didn't understand me" | Follow-ups, Nudges, Prompt Enhancer |
| "I want alternatives" | Variations, Branches, Randomize |
| "Is this accurate?" | Citations, References, Caveat |
| "This is taking too long" | Draft Mode, Controls, Cost Estimates |
| "I need AI to do something complex" | Action Plan, Stream of Thought |
| "Is this AI or human?" | Disclosure, Avatar, Name |
| "Don't store my data" | Incognito Mode, Consent, Data Ownership |
| "AI forgot what I said" | Memory (scoped/global/ephemeral) |
By AI Product Type
| Product Type | Essential Patterns | Nice-to-Have |
| Chat assistant | Open Input, Suggestions, Follow-ups, Memory, Disclosure | Gallery, Voice & Tone, Branches |
| Code copilot | Inline Action, Stream of Thought, Controls, Citations | Action Plan, Draft Mode |
| Image generator | Gallery, Parameters, Variations, Inpainting, Preset Styles | Draft Mode, Randomize |
| Document AI | Attachments, Citations, Caveat, Disclosure, Summary | Transform, Expand, Follow-ups |
| AI agent | Action Plan, Controls, Verification, Stream of Thought, Cost Estimates | Memory, Consent |
| Voice assistant | Voice Avatar, Personality, Controls, Disclosure | Memory, Consent |
| Enterprise analytics | Citations, Connectors, Filters, Modes, Disclosure | Action Plan, Memory |
Trust Level Decision
High-stakes domain (healthcare, finance, legal)?
YES -> CRITICAL: Citations + Verification + Disclosure + Caveat + Audit
NO -> AI output mixed with human content?
YES -> HIGH: Disclosure + Citations + Caveat
NO -> Could AI output cause harm if wrong?
YES -> MEDIUM: Caveat + Citations (optional)
NO -> LOW: Minimal caveat, focus on UX qualityThe Six Pattern Categories
1. Wayfinders — Help users construct their first prompt
Gallery, Suggestions, Templates, Follow-ups, Initial CTA, Nudges, Prompt Details, Randomize
2. Prompt Actions — Actions users direct AI to complete
Open Input, Inline Action, Chained Action, Regenerate, Transform, Restyle, Expand, Summary, Synthesis, Describe, Auto-fill, Restructure, Madlibs, Inpainting
3. Tuners — Adjust context and settings to refine prompts
Attachments, Connectors, Parameters, Model Management, Modes, Filters, Prompt Enhancer, Preset Styles, Saved Styles, Voice and Tone
4. Governors — Human-in-the-loop oversight and agency
Action Plan, Stream of Thought, Controls, Draft Mode, Branches, Variations, Citations, References, Verification, Memory, Cost Estimates, Sample Response, Shared Vision
5. Trust Builders — Confidence in AI ethics and accuracy
Disclosure, Caveat, Consent, Data Ownership, Watermark, Footprints, Incognito Mode
6. Identifiers — Distinct AI qualities at brand/model level
Avatar, Personality, Name, Color, Iconography
CRITICAL Gotchas
| Rule | Why |
| NEVER use photorealistic avatars unless AI matches that capability | Sets unrealistic expectations, erodes trust |
| ALWAYS show Stream of Thought for tasks > 5 seconds | Users abandon when they can't see progress |
| NEVER let Memory be a black box | Users must see, edit, and delete what AI remembers |
| ALWAYS offer Controls (at minimum: stop) during generation | Users need escape hatches |
| NEVER rely solely on Caveats for safety | Caveat blindness is real; design the system to be safe |
| ALWAYS distinguish AI content from human content in blended UIs | Users may unknowingly present AI work as their own |
| NEVER overwrite user work without Verification | Accidental overwrites destroy trust instantly |
Related Skills
- 21-enterprise-ai-ux - Enterprise-specific AI design
- 22-conversation-ux - Thread management and conversation data models
- 23-uiux-design-principles - General design principles
- 20-interactive-widgets - Widget rendering and state management
Support
uiux-designer- AI-specific interaction pattern selection and designkaizen-specialist- AI agent capabilities informing UX decisionsreact-specialist- Implementation of AI interaction patterns
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.