Understanding Processing Statuses & OCR Confidence

When you upload a document to DocuNero, it moves through a series of processing stages while OCR and AI extraction work to interpret your data. Understanding these statuses—and the confidence scores assigned to each field—helps you quickly identify which documents need review and which are ready for export.

Processing Pipeline

DocuNero's extraction process uses Google Vision OCR and AI to identify vendors, dates, amounts, and line items with confidence scoring for each field.

1. Processing Statuses in DocuNero

Every document goes through several stages during extraction. These appear in your Upload Center, Documents page, and Batch overview.

StatusDescriptionWhat to Do
UploadedThe document has been successfully uploaded to DocuNero.Processing will begin automatically.
ProcessingThe document is actively being analyzed by DocuNero's OCR + AI pipeline.Processing usually finishes in under a minute.
Processed/ParsedExtraction was successful and all available fields were identified.Review and approve the extracted data.
Error/FailThe system was unable to extract data from the document.Try re-uploading or check file clarity.

Uploaded

The document has been successfully uploaded to DocuNero's servers.

What happens next: The system will automatically begin processing the document.

What to do: Wait for the status to change to "Processing".

Processing

What happens during this stage:

  • Google Vision OCR extracts text, layout, and visual structure
  • AI identifies vendors, dates, totals, taxes, and line items
  • Confidence scores are calculated
  • Data is structured for review

Duration: Typically 10–40 seconds, depending on clarity and complexity.

Processed/Parsed

Extraction was successful! The document includes:

  • Vendor name
  • Dates
  • Subtotal, tax, and totals
  • Line items
  • Payment method (if available)
  • Category suggestions (auto)

Next step: Review and approve the extracted data.

Error/Fail

The system was unable to extract data from the document.

Common reasons:

  • Very blurry or dark images
  • Handwritten receipts
  • Scans with missing sections
  • Encrypted or corrupted PDFs
  • Unsupported file types (rare)

How to fix: Try re-uploading, use higher-quality images, or contact support.

2. Understanding OCR Confidence Scores

DocuNero assigns each extracted field a confidence score, showing how certain the AI is about the value.

Confidence is displayed as a percentage with colored indicators and tooltips explaining the score.

High Confidence (90–100%)

The extracted value is almost certainly correct.

Examples:

  • Clear printed text
  • High-resolution PDFs
  • Items with strong formatting (invoice totals, dates)

Recommended action: Quick glance review only.

Medium Confidence (70–89%)

OCR is reasonably confident, but verification is suggested.

Common causes:

  • Slight blur
  • Faded text
  • Slight skew or shadow

Recommended action: Review the value before approving.

Low Confidence (Below 70%)

OCR is unsure about the extracted value.

Common causes:

  • Poor lighting
  • Handwritten fields
  • Cropped or cut-off text
  • Noisy or cluttered background

Recommended action: Manually correct the field.

3. How Confidence Scores Affect Your Workflow

Prioritize documents
Review low-confidence items first
Reduce manual work
High-confidence documents need minimal review
Improve AI accuracy
Correcting fields helps train the model

4. Best Practices for Improving OCR Confidence

Quality Matters

Higher quality uploads = higher confidence scores = faster processing and fewer corrections needed.

Do:

  • ✔ Upload PDFs when possible
  • ✔ Take photos in bright, even lighting
  • ✔ Ensure the document is flat and centered
  • ✔ Increase resolution when scanning (200–300 DPI recommended)

Avoid:

  • ❌ Dark photos or shadows
  • ❌ Angled or skewed images
  • ❌ Crumpled or folded receipts
  • ❌ Screenshots of screenshots

5. When to Contact Support

Reach out to support if:

  • Many documents repeatedly fail
  • Confidence scores are consistently low across high-quality uploads
  • A specific vendor's invoices frequently extract incorrectly
  • You need higher upload limits for bulk processing

DocuNero's support team can assist with troubleshooting and improving your workflow.

6. Summary

Understanding document statuses and OCR confidence helps you:

  • Track processing progress
  • Spot issues early
  • Speed up review and approval
  • Ensure the highest possible data accuracy

DocuNero's extraction pipeline is optimized to give you clear visibility into every step of the process.

Frequently Asked Questions

How long does document processing typically take?

Processing usually completes in 10–40 seconds, depending on document clarity, complexity, and current system load.

What should I do if a document fails to process?

Try re-uploading with higher quality, ensure the document isn't password-protected, and check that text is legible. Contact support for persistent issues.

How accurate are the OCR confidence scores?

Confidence scores reflect the AI's certainty level. High confidence (90%+) are usually correct, while low confidence (<70%) need manual verification.

What happens if I approve a document with low confidence scores?

You can still export the data, but we recommend verifying low-confidence fields first to ensure accuracy in your records.