Tag Suggestions

Automatically categorize tickets with AI-suggested tags. Keep your ticket organization consistent without manual tagging effort.

What Are Tag Suggestions?

Tag suggestions use AI to analyze ticket content and recommend relevant tags from your organization's tag taxonomy.

Example

Ticket: "I need a refund for order #12345. The product arrived damaged."

Suggested tags:

• refund
• damaged-product
• shipping-issue

How Tag Suggestions Work

Analysis Process

1. AI reads ticket subject and description
2. Matches content against your existing tags
3. Suggests 3-5 most relevant tags
4. Shows confidence for each suggestion

What AI Considers

• Keywords in the ticket
• Context and meaning
• Your tag taxonomy
• Historical tag usage
• Similar ticket patterns

Where to See Suggestions

AI Sidebar

View suggested tags in the AI Insights panel:

🏷️ Suggested Tags
• refund (95%)
• damaged-product (88%)
• shipping-issue (72%)

Ticket Creation

When creating tickets manually, suggestions appear as you type.

Ticket Detail

Suggested tags show below current tags with "Add" buttons.

Using Tag Suggestions

Adding Suggested Tags

1. View suggestions in sidebar
2. Click a tag to add it
3. Or click "Add all" for bulk add

Dismissing Suggestions

• Click ✕ on a suggestion to dismiss
• Dismissed tags won't be suggested again for this ticket

Requesting New Suggestions

Click "Refresh" to regenerate suggestions based on updated content.

Setting Up for Best Results

Build Your Tag Taxonomy

Good suggestions require good tags. Create tags for:

Categories:

• billing, shipping, product, account

Issue types:

• refund, exchange, complaint, question

Products/services:

• product-a, product-b, subscription

Sentiment/priority:

• vip, escalated, positive-feedback

Tag Naming Best Practices

Automation with Tags

Auto-Apply Suggested Tags

In Settings → AI → Tags:

• Enable auto-apply
• Set confidence threshold (e.g., > 85%)
• Tags applied automatically when confident

Tag-Based Routing

Combine with automations:

IF tag contains "billing"
THEN assign to billing-team

IF tag contains "vip"
THEN set priority = High

Tag Accuracy

Improving Suggestions

1. Use consistent tags: Same tag for same concepts
2. Descriptive tag names: "damaged-product" not "dp"
3. Regular cleanup: Remove unused tags
4. Merge duplicates: "refund" and "refunds" → "refund"

Confidence Levels

Managing Your Tag System

Viewing All Tags

Go to Settings → Tags to see:

• All organization tags
• Usage counts
• Creation date
• Last used

Creating Tags

1. Go to Settings → Tags
2. Click "Add Tag"
3. Enter name and optional description
4. Add color (optional)
5. Save

Merging Tags

When you have duplicates:

1. Go to Settings → Tags
2. Select tags to merge
3. Choose primary tag
4. Click "Merge"
5. All tickets updated automatically

Deleting Tags

1. Go to Settings → Tags
2. Select tag to delete
3. Click "Delete"
4. Confirm removal from all tickets

Best Practices

For Organization

1. Start simple: Begin with 10-20 core tags
2. Expand thoughtfully: Add tags as patterns emerge
3. Review regularly: Prune unused tags
4. Document meanings: Clear tag descriptions

For Using Suggestions

1. Review before adding: Don't blindly accept
2. Correct mistakes: AI learns from corrections
3. Use consistently: Same situation = same tags
4. Trust high confidence: 90%+ is usually right

Tag Analytics

Viewing Tag Metrics

In Analytics → Tags:

• Most used tags
• Tag trends over time
• Tags by ticket volume
• Resolution time by tag

Useful Insights

• Spike in "bug" tags: Product issue?
• Rising "billing" tags: Pricing confusion?
• "Escalated" trends: Support quality issue?

Troubleshooting

"No tag suggestions"

• Check AI settings are enabled
• Verify you have tags created
• Ticket may be too short

"Wrong tag suggestions"

• Your tags may need clearer names
• AI learns from corrections
• Review your tag taxonomy

"Same tags suggested for everything"

• Tags may be too broad
• Add more specific tags
• Check ticket content variety

"Tag suggestions are slow"

• Complex tickets take longer
• Check connection
• Try refreshing

← Priority Prediction | Article Recommendations →