How Confidence Scoring Works

Every field CargoLint extracts comes with a confidence score—a number between 0 and 1 that represents how certain the AI is about its answer. Understanding these scores helps you know which extractions to trust and which need a closer look.

What the confidence scale means

• 0.9–1.0: Excellent confidence. The AI found clear, unambiguous data that matches expected patterns. These extractions rarely need review.
• 0.7–0.9: Good confidence. The data was found with reasonable certainty, but there are minor ambiguities or variations from the expected format. Most of these pass without review.
• 0.5–0.7: Moderate confidence. The data was found but with some uncertainty—possibly due to unclear text, unexpected layout, or multiple possible interpretations. These typically go to review.
• 0.0–0.5: Low confidence. The AI struggled to find or interpret the data. These always go to review.

What affects confidence scores

Document quality and clarity

• Clean, high-resolution scans produce higher scores than blurry or low-contrast images.
• Handwritten text typically scores lower than printed text because it’s harder for the AI to parse.
• Poor lighting, shadows, or damage to the document reduces confidence.

Field location and format consistency

• Fields in expected locations (e.g., invoice number near the top) score higher.
• Consistent formatting (dates in the same format, amounts with the same currency symbol) increases confidence.
• Unusual layouts or fields in unexpected places lower scores.

Text patterns and validation

• The AI validates extracted data against patterns. For example, if an invoice total should equal the sum of line items, a mismatch reduces confidence.
• Fields that match known patterns (valid dates, valid currencies, properly formatted addresses) score higher.
• Incomplete or malformed data (missing ZIP codes, invalid dates) lowers confidence.

Language and character recognition

• Text in the primary language of the document scores higher.
• Documents with mixed languages or unusual characters may have lower confidence on those sections.

Special penalties

CargoLint applies confidence adjustments for specific issues:

• Missing required field: -0.20 (The field should be there but wasn’t found)
• Calculation mismatch: -0.20 (Extracted values don’t match expected math, like totals)
• Format warning: -0.10 (Data is present but in an unexpected format)

These penalties stack, so a field with multiple issues can see a significant confidence drop.

The 70% review threshold

CargoLint automatically routes documents to your review queue when the overall confidence score falls below 70%. This threshold balances accuracy with speed—high-confidence documents process instantly, while lower-confidence ones get human attention.

You can adjust this threshold in your workspace settings if your workflow demands higher or lower accuracy.

Per-document-type thresholds

Different document types may have slightly different confidence thresholds because they have different complexity levels:

• Commercial Invoices: 70% baseline (highly standardized)
• Bills of Lading: 70% baseline (highly standardized)
• Packing Lists: 70% baseline (moderately variable)
• Certificates of Origin: 70% baseline (highly variable by region)

How corrections improve future scores

When you review and correct an extraction in CargoLint, the system learns from your corrections. Over time, corrections to specific field types and document layouts improve the AI’s confidence on similar documents, which can lower the proportion of documents requiring review.

This learning is continuous—the more you use CargoLint, the smarter it gets about your documents.