AI document classification
How ImmPilot's classifier identifies uploaded documents and extracts key fields.
When a document is uploaded, the classifier examines the content and predicts the document type from a taxonomy of over 200 immigration-relevant categories. It also extracts a set of key fields appropriate to the predicted type.
Confidence and confirmation
Every classification comes with a confidence score. High-confidence classifications are surfaced as suggestions for a team member to confirm with one click. Low-confidence classifications require manual classification. The classifier improves against your firm's confirmations over time.
Extracted fields
For each document type, the classifier extracts the fields most useful for case management — passport numbers and expiries, ECA scores, employment dates, education levels. Extracted fields are validated by the team member during confirmation.
Privacy and data handling
Documents are processed by our AI provider under a strict data processing agreement. No document content is used to train external models. All processing occurs within Canadian data residency where applicable.