PRs #20-21 Feature Testing
Time: 10-15 minutes
PR #20 - Clarity Assistant:
- AI checks your message before you send it
- Suggests improvements for clarity
- Warns about tone issues
- Shows alternative phrasing
PR #21 - Action Items:
- Auto-extracts tasks from conversations
- Track who needs to do what
- Mark items complete
- Filter by status
Clarity checking happens automatically while typing!
In any chat, start typing:
- Type: "can u do the thing tmrw?"
- Wait 2 seconds (debounced)
- Look for clarity suggestion to appear
Expected:
- Clarity suggestion view appears
- Shows issues: "Unclear reference", "Informal language"
- Suggests: "Can you complete [task] by tomorrow?"
- Options: Accept | Dismiss | More
✅ Pass if:
- Suggestion appears after typing
- Issues clearly explained
- Can accept or dismiss
- Accepting replaces your text
Type a clear message:
- Type: "I will complete the report by Friday at 5pm"
- Wait 2 seconds
Expected:
- NO suggestion appears
- Message is clear, no issues
- Can send directly
✅ Pass if:
- No unnecessary suggestions
- Clear messages pass through
Type something with wrong tone:
- Type: "This is totally wrong and you messed up!"
- Wait 2 seconds
Expected:
- Suggestion appears
- Tone warning: "Message may sound aggressive"
- Suggests softer alternative
- Can accept or dismiss
✅ Pass if:
- Tone issues detected
- Helpful suggestions provided
Action items extracted from conversation automatically
From a conversation:
- Look for action items button (should be in toolbar)
- Or check conversation list for action item indicators
Expected:
- Access point to ActionItemsView
- Can see list of extracted items
Open ActionItemsView:
Expected to see:
- Filter tabs: All | Open | Completed
- List of action items extracted from messages
- Each item shows:
- Description
- Assignee (if mentioned)
- Due date (if mentioned)
- Checkbox to mark complete
✅ Pass if:
- Items displayed clearly
- Can switch between filters
- UI is clean and organized
Tap checkbox on an action item:
Expected:
- Checkmark appears
- Item marked complete
- Moves to "Completed" tab
- Visual indication (strikethrough or grey)
✅ Pass if:
- Completion toggles work
- Filters update correctly
- State persists
Swipe left on an action item:
Expected:
- Red delete button appears
- Tap to delete
- Item removed from list
✅ Pass if:
- Swipe gesture works
- Deletion works
- No crash
In a conversation:
- Try typing (clarity check)
- Tap summarize (AI summary)
- View action items
- Send messages (real-time)
✅ Pass if:
- All features work together
- No conflicts
- No crashes
- Smooth experience
Have this conversation:
Alice: "Can you review the PR by Friday?"
Bob: "Yes, I'll review it by Friday 3pm"
Then:
- Extract action items (if auto-extract enabled)
- Or manually trigger extraction
Expected:
- Action item appears: "Review PR by Friday 3pm"
- Assigned to: Bob
- Due: Friday
✅ Pass if:
- AI correctly extracts commitment
- Assignee identified
- Deadline captured
- Suggestions appear: ☐ Pass ☐ Fail
- Issues identified correctly: ☐ Pass ☐ Fail
- Accept/Dismiss works: ☐ Pass ☐ Fail
- Debouncing works (2 sec delay): ☐ Pass ☐ Fail
- View accessible: ☐ Pass ☐ Fail
- Items display correctly: ☐ Pass ☐ Fail
- Mark complete works: ☐ Pass ☐ Fail
- Filters work: ☐ Pass ☐ Fail
- Swipe-to-delete works: ☐ Pass ☐ Fail
- Features work together: ☐ Pass ☐ Fail
- No crashes: ☐ Pass ☐ Fail
Mock AI Responses:
- Clarity and Action Items may return mock/empty responses
- This is OK for testing the UI/UX
- Cloud Functions need real OpenAI integration
Expected Behavior:
- Clarity: May not show suggestions (mock returns empty)
- Actions: May not find items (mock returns empty)
- Summarization: Works with real GPT-4!
This tests the infrastructure, not the AI quality yet.
UI/UX works: Features accessible, no crashes
Integration works: All features coexist
Code quality: Clean, professional
AI Quality: To be tested when functions fully integrated
Start testing! Report any issues you find. 🧪