Skip to main content
Duplicate Detection uses AI to identify when multiple tickets are about the same issue. This helps your team avoid duplicate work and provides consistent answers.
Duplicate Detection is a Pro feature. Upgrade your plan to access it.

How It Works

  1. User opens a new ticket
  2. AI compares it to recent tickets (same user and others)
  3. If a similar ticket is found, staff is notified
  4. Staff can link or merge the tickets
Example:
User opens Ticket #42: “My payment isn’t going through” User opened Ticket #40 yesterday: “Payment failing on checkout” AI detects these are likely the same issue and alerts staff.

Why It Matters

Avoid Duplicate Work

Without detection, two staff members might work on the same issue independently, wasting time and potentially giving conflicting answers.

Consistent Answers

When tickets are linked, staff can see previous responses and provide consistent information.

Faster Resolution

If a similar ticket was already resolved, staff can apply the same solution immediately.

Getting Started

1

Enable Duplicate Detection

Go to AI FeaturesDuplicate Detection and toggle it on.
2

Configure Thresholds

Set how confident AI should be before flagging duplicates.
3

Test It

Open two tickets with similar content to verify detection works.

Confidence Levels

AI assigns a confidence level to potential duplicates:
LevelConfidenceWhat It Means
Very High90%+Almost certainly the same issue
High75-90%Very likely related
Medium60-75%Possibly related, review recommended
Low45-60%Might be related, less certain
Start with High threshold (75%) and adjust based on your experience. Too low = false positives. Too high = missed duplicates.

Settings

Confidence Thresholds

Customize the threshold for each confidence level:
Very High: 90%+    → Automatically notify
High: 75-90%       → Flag for review
Medium: 60-75%     → Suggest possible match
Low: 45-60%        → Only show on request

Time Window

How far back to search for duplicates:
  • 24 hours - Only very recent tickets
  • 7 days - Recommended for most teams
  • 30 days - For recurring issues

Scope

What tickets to compare against:
OptionDescription
Same User OnlyOnly compare tickets from the same user
All UsersCompare against all recent tickets
Same CategoryOnly compare within the same category

How Detection Works

Content Similarity

AI compares the meaning, not just exact words:
Ticket A: "My payment isn't working"
Ticket B: "Payment failed at checkout"
→ Detected as similar (same meaning)

Topic Matching

AI understands the general topic:
Ticket A: "How do I upgrade my plan?"
Ticket B: "What's the process for upgrading?"
→ Detected as similar (same topic)

User Context

The same user opening multiple tickets is weighted higher:
Same user + similar content = likely duplicate
Different users + similar content = might be same issue affecting multiple users

What Staff Sees

When a duplicate is detected, staff sees:
  1. Notification in the ticket with confidence level
  2. Link to the potential duplicate ticket
  3. Options to:
    • View the other ticket
    • Link the tickets
    • Dismiss the suggestion

Actions

Connect two related tickets without merging them:
  • Both tickets remain open
  • Staff can see they’re related
  • Good for related but distinct issues

Dismiss

Mark the suggestion as incorrect:
  • AI learns from this feedback
  • The suggestion won’t appear again
  • Use when tickets aren’t actually related

Best Practices

Review, Don’t Auto-Merge

Duplicate detection suggests matches - staff should verify:
  • Are they truly the same issue?
  • Should they be linked or merged?
  • Is additional context needed?

Use Consistent Subject Lines

Help AI by using clear subject lines:
  • ✅ “Payment failing with error code 401”
  • ❌ “Help!!!” (hard to match)

Check Before Creating

Encourage users to check existing tickets:
  • Reduces actual duplicates
  • Makes detection more accurate
  • Improves support efficiency

Troubleshooting

Too Many False Positives

If unrelated tickets are being flagged:
  1. Raise threshold - Increase confidence requirement
  2. Narrow scope - Check “Same User Only” or “Same Category”
  3. Reduce time window - Shorten to 24-48 hours

Missing Obvious Duplicates

If duplicates aren’t being detected:
  1. Lower threshold - Decrease confidence requirement
  2. Expand scope - Check “All Users”
  3. Extend time window - Increase to 14-30 days

Same User, Different Issues

If different issues from the same user are flagged:
  1. Check thresholds - Make sure “Medium” isn’t too low
  2. Review scope - “Same Category” may help
  3. Train your users - Encourage clear subject lines

Need Help?

Get instant support via private DM