UX Waiting States Audit

Audit how web applications handle long-running operations (30+ seconds) using AI-powered browser automation.

Built while improving wait time UX at onsa.ai — a multi-agent B2B sales platform where AI agents take 30-60+ seconds to perform real research.

The Problem

AI agent products often have long wait times. Users feel anxious, abandon operations, or perceive the product as "slow" — even when the wait is justified.

The Psychology of Waiting

Research shows 5 key factors affect perceived wait time:

1. Occupied time feels shorter than unoccupied time (the "watched pot" effect)
2. Known waits feel shorter than uncertain waits
3. Explained waits feel shorter than unexplained waits
4. Calm waits feel shorter than anxious waits
5. Group waits feel shorter than solo waits

The 10-Point Checklist

This audit evaluates:

#	Category	Question
1	Progressive Value	Does user see partial results early?
2	Heartbeat Indicators	Are there signs system is actively working?
3	Time Estimation	Does user know how long to wait?
4	Process Explanation	Does user understand why it takes time?
5	Sunk Cost Visibility	Does user see accumulated work?
6	Work While Waiting	Can user do other tasks meanwhile?
7	Interruptible	Can user get "good enough" results early?
8	Graceful Degradation	Does partial failure still show results?
9	Completion Celebration	Is the ending a positive moment?
10	Anxiety Reduction	Is it clear system hasn't frozen?

Installation

Claude Code Skill

1. Clone this repo or download the files

2. Add the skill to Claude Code:

   • Copy folder to ~/.claude/skills/ux-waiting-audit/
   • Or import ux-waiting-audit.skill via Claude Code settings

3. Invoke with: "Run a UX waiting audit on [URL]"

Manual Use

Use the checklist in references/checklist.md to manually evaluate any application.

Files

File	Description
SKILL.md	Main skill definition for Claude Code
references/checklist.md	Detailed 10-point evaluation criteria with detection scripts
references/report-template.md	Template for generating audit reports
scripts/capture_state.js	JavaScript helpers for detecting UI states

Usage Example

Audit the waiting UX on https://app.example.com

1. Navigate to the dashboard
2. Trigger a report generation (takes ~45 seconds)
3. Evaluate against the UX waiting checklist
4. Generate a report with screenshots and recommendations

Sample Output

The audit generates:

• Screenshots timeline at T+0s, T+10s, T+30s, T+Complete
• Score: X/10 checklist items addressed
• Strengths: What works well
• Critical gaps: Missing elements hurting UX
• Quick wins: Low-effort, high-impact fixes
• Priority matrix: P1/P2/P3 recommendations

Best-in-Class Examples

Reference implementations of excellent waiting UX:

• Figma exports: Progress bar + percentage + file count
• Notion AI: Streaming text + cursor animation
• ChatGPT: Token-by-token streaming + stop button
• Linear search: Instant partial results + refinement
• Vercel deployments: Step-by-step progress + logs

Credits

• Built by Bayram Annakov
• Inspired by UX research at onsa.ai

License

MIT