A U.S. death certificate contains over 30 distinct data fields. The core identifying fields — legal name, Social Security number, date of birth, date of death, and state of death — are required by virtually every institution that receives a notification letter. Beyond those, a certificate typically includes the place of birth, race, sex, occupation, the cause and manner of death, the place where the death occurred, and the name and relationship of the informant. Every field matters: institutions cross-reference the information in notification letters against their own records, and discrepancies cause delays.
Every U.S. state issues its own death certificate on its own form, which means field layouts, label names, and document structures vary considerably across jurisdictions. A California certificate from 2024 differs from a Texas certificate from the same year — and both differ from certificates issued a decade ago. Handwriting, scan quality, and watermarks add further complexity. This is why accurate automated extraction requires a model that understands the document spatially, not just character by character. Sedare's extraction pipeline is tested on real-world certificates from 14 states and validates every field against 40+ rules before presenting results for review.
Sedare handles estate notification letters in minutes, not months.
Get Early Access