Priority Prediction
Let AI help you prioritize. Priority prediction analyzes tickets and suggests the appropriate urgency level.
What Is Priority Prediction?
Priority prediction uses AI to recommend a ticket's urgency based on:
- Message content and tone
- Customer sentiment
- Issue type and impact
- Business context
- Time-sensitive language
Priority Levels
| Priority | Description | Typical Response |
|---|---|---|
| Urgent 🔴 | Critical business impact, time-sensitive | < 1 hour |
| High 🟠 | Significant impact, needs attention | < 4 hours |
| Normal 🔵 | Standard request | < 24 hours |
| Low ⚪ | Minor issue, when time permits | < 48 hours |
How Prediction Works
Signals AI Looks For
Urgent indicators:
- "System is down"
- "Can't process orders"
- "Security breach"
- "ASAP", "immediately", "urgent"
- All caps, multiple exclamation marks
- Deadline mentions
High priority indicators:
- "Affecting multiple users"
- "Losing money/customers"
- "Been waiting a week"
- Frustrated sentiment
- VIP customer markers
Normal indicators:
- Standard questions
- Feature requests
- General inquiries
- Neutral sentiment
Low priority indicators:
- "When you get a chance"
- "No rush"
- Minor suggestions
- Informational queries
Where to See Predictions
AI Sidebar
Suggested priority appears in the AI Insights panel with confidence score.
New Ticket Creation
When tickets are created, AI suggests priority automatically.
Ticket List
Priority badge reflects current priority (either AI-suggested if new, or manually set).
Using Priority Predictions
Accept Suggestion
If AI's prediction seems right:
- View the suggested priority
- Click "Accept" or leave as-is
- Priority is set
Override Prediction
If you disagree:
- Click the priority badge
- Select correct priority
- Your choice is saved
Batch Review
For triage:
- View tickets sorted by AI-suggested priority
- Review urgent suggestions first
- Adjust as needed
Factors That Influence Predictions
Customer Factors
- VIP status: Known VIP customers may get higher suggestions
- History: Customers with past urgent issues
- Sentiment: Negative sentiment increases priority
Content Factors
- Keywords: Time-sensitive language
- Issue type: Known critical issue types
- Impact scope: Individual vs. multiple users
Business Factors
- Business hours: After-hours may adjust urgency
- Similar tickets: Spike in similar issues
- Integration data: Order value, subscription tier
Configuring Priority Prediction
Organization Settings
In Settings → AI → Priority:
- Auto-apply: Automatically set predicted priority (vs. suggest only)
- Confidence threshold: Only apply if confidence > X%
- Default priority: When prediction is uncertain
- VIP boost: Auto-increase priority for VIP customers
Priority Rules
Combine AI predictions with rules:
IF predicted priority = Urgent
AND customer_tier = Enterprise
THEN set priority = Urgent AND notify @on-call
Accuracy & Limitations
Generally Accurate For
- Clear urgency language
- Standard support scenarios
- Known issue types
- Established patterns
May Misjudge
- Context-dependent urgency
- Internal business priorities
- Customer relationship nuances
- Edge cases
Continuous Improvement
AI learns from:
- Your priority adjustments
- Resolution patterns
- Feedback on predictions
Best Practices
Trust but Verify
- Check AI suggestion
- Apply your judgment
- Adjust if needed
- AI improves from your corrections
Don't Over-Prioritize
- Not everything is urgent
- Urgency inflation helps no one
- Be consistent with priority meanings
Regular Calibration
- Review priority accuracy monthly
- Adjust settings as needed
- Ensure team uses priorities consistently
Priority-Based Workflows
SLA Integration
Priorities typically map to SLA targets:
| Priority | First Response | Resolution |
|---|---|---|
| Urgent | 1 hour | 4 hours |
| High | 4 hours | 24 hours |
| Normal | 24 hours | 72 hours |
| Low | 48 hours | 1 week |
Automation Examples
Auto-assign urgent tickets:
Trigger: Priority = Urgent
Action: Assign to on-call agent
Escalate aging high priority:
Trigger: Priority = High AND age > 4 hours AND unassigned
Action: Notify manager
Troubleshooting
"Prediction seems wrong"
- AI may lack context you have
- Override and continue
- AI will learn from the correction
"All tickets predicted as normal"
- Check AI settings
- Review recent ticket content
- May need more urgency signals
"Priority keeps changing"
- AI updates as conversation evolves
- Major sentiment shifts may affect priority
- Manual overrides are preserved