Extraction Failure Logging

Explanation

Extraction failure logging checks that whenever a document-extraction system fails to capture a field — leaving it blank, unreadable, or rejected — that failure is recorded rather than silently passing an incomplete record downstream. It matters in BFSI because a missed field on an onboarding form or loan document can break KYC, misstate a record, or stall processing, and without logging nobody can see how often or where extraction is breaking down. It is measured by the Field Extraction Failure Rate: the proportion of expected fields the system fails to extract. No explicit formula is given, so report the failure rate over attempted field extractions without inventing additional terms. Implement it by instrumenting the extraction stage to emit a structured log entry for each failed field, including the document id, field name, and failure reason, and aggregate these into a monitored rate that forms the audit evidence trail. The target is a failure rate of less-than-or-equal-to 1% (ideally 0); any breach raises an alert so operators can investigate the cause — a degraded parser, a new document layout, poor scan quality — and remediate before more records are processed with missing data.