Handling Review Items
What to do when a submission is flagged as "Needs Review" and how to resolve it quickly.
Handling Review Items
Sometimes Doculent will flag a submission as Needs Review. This isn't a failure — it means the system found something it wants a human to verify before moving forward.
Think of it as Doculent raising its hand and saying "I'm not 100% sure about this one — can you take a look?"
Why Does a Submission Need Review?
The most common reasons:
| Reason | Example |
|---|---|
| Low confidence extraction | A handwritten field that was hard to read |
| Missing required field | The document didn't contain an expected piece of data |
| Unusual value | A date that seems way off, or a number that's outside the expected range |
| Validation rule failure | The extracted data didn't pass one of your configured quality checks |
Finding Items That Need Review
You can quickly find all items waiting for your attention:
- Go to Submissions
- Filter by status: Needs Review
You'll see a list of all submissions that need your input, with the most recent at the top.
Make it a habit to check the Needs Review queue regularly — once in the morning and once in the afternoon works well for most teams.
Resolving a Review Item
Open the submission and you'll see the flagged fields highlighted. For each one:
- Look at the flag reason — it'll tell you why this field was flagged (low confidence, validation failure, etc.)
- Check the original document — the source document is displayed right next to the extracted data for easy comparison
- Take action:
- Confirm the value — if the extraction is actually correct, click Approve to accept it as-is
- Edit the value — if it's wrong, click the field, type the correct value, and save
- Reject the submission — if the document is unprocessable or irrelevant
Once all flagged fields are resolved, the submission moves forward to delivery automatically.
Bulk Review
If you have many items to review, you can work through them efficiently:
- Use the Next button to move through review items one by one without going back to the list
- Common patterns emerge quickly — if the same field is flagged repeatedly, consider adjusting your validation rules to be more lenient (or fixing the upstream document quality)
Reducing Review Items Over Time
Too many items ending up in review? Here are some things to try:
| Problem | Solution |
|---|---|
| Low confidence on specific fields | Check if source documents are high quality (clear scans, not photos) |
| Same field always flagged | Adjust the confidence threshold in Channel Settings |
| Validation too strict | Loosen the rules in Validation Rules |
| Wrong document types coming in | Tighten Classification Rules to reject irrelevant docs |
The goal is to get your review queue as small as possible. A well-configured channel should auto-approve the vast majority of submissions, leaving you with only the genuinely ambiguous cases.