Using the Expense Review Queue

Using the Expense Review Queue

The Processing Queue helps you manage AI-processed expenses that need review or encountered issues. Monitor processing status and handle failed extractions efficiently.

Accessing the Review Queue

1. Go to Expenses in the main navigation
2. Look for the Review Queue button
3. Red badge shows count of items needing attention
4. Click to open the Processing Queue page

Understanding Queue Statistics

The dashboard displays four key metrics:

• Processing Count - Items currently being analyzed (spinner icon)
• Queued - Waiting to be processed (clock icon)
• Failed - Processing errors that need retry (X icon)
• Success Rate - Percentage of successful extractions

Currently Processing Section

• Shows active AI analysis in real-time
• Displays processing time elapsed
• Shows attempt count for retries
• Format: "Processing for 2m 15s - Attempt 1"

Queued for Processing

• Documents waiting in line
• Displayed with clock icon badge
• Automatically processes when resources available
• Shows upload time and file name

Failed Processing Section

• Items that encountered errors
• Shows error type label
• Retry button available for single items
• Retry All button for bulk retry of all failed items

Reviewing Individual Expenses

Click on any processed expense to review:

• View original receipt image
• Check extracted data fields
• See confidence scores per field
• Edit any incorrect information
• Approve or reject the expense

Common Failure Reasons

• Image Quality - Blurry or low-resolution photos
• Unreadable Text - Faded receipts or poor lighting
• Invalid Format - Corrupted files or unsupported types
• Missing Information - Receipt lacks required fields

Auto-Refresh Toggle

Monitor processing in real-time:

• Enable auto-refresh to watch progress live
• Page updates every few seconds
• See status changes as they happen
• Useful when processing large batches

Best Practices

• Check review queue daily after bulk uploads
• Address failed items promptly
• Re-upload poor quality images for better results
• Use manual entry for consistently failing documents