What to do
Keep an ambiguous numeric date unchanged until evidence resolves its meaning. A standard output format can express a decision, but cannot make that decision for you. Record the original text, supporting evidence, and reviewed result together.
Supplier address, browser language, and file metadata alone may not establish the intended date convention.
Two worked review cases
These cases are fictional teaching data, not measured extraction results. Use the source excerpts to practise deciding what belongs in the reviewed output.
| Case | Fictional source excerpt | Review decision |
|---|---|---|
| Ambiguous source | 04/05/2026 with no supporting format detail | Hold; both April 5 and 4 May remain possible. |
| Explicit source | 4 May 2026 | A reviewed year-month-day value can be 2026-05-04. |
Suggested review columns
| Column | Definition or instruction |
|---|---|
| Raw date | Exact printed date string and its label. |
| Evidence for interpretation | Explicit month wording or confirmed supplier clarification. |
| Reviewed date | Normalized value only after the ambiguity is resolved. |
These are suggested staging-sheet columns. Printed facts can inform field descriptions in Parzable’s Schemas view; source references, review statuses, and decisions must be supplied by your own review process.
A practical review sequence
- Locate the date label so issue and due dates are not confused.
- Preserve the raw string and identify evidence needed for interpretation.
- Record the confirmed meaning before importing as a date type.
Errors to avoid
- Do not let an automatic import silently resolve the ambiguity.
- A plausible due-date interval is supporting context, not conclusive evidence by itself.
Use the case worksheet with a small sample
Download this page’s worked-case worksheet. It contains the two fictional cases above and a blank row for your own review exercise. Keep any real source documents in your agreed storage process.
For the broader product workflow, see the related guide. Parzable accepts PDF/image inputs and downloads CSV. Suggested structures here are review targets; nested item rows, automatic approvals, and destination integrations are not guaranteed.
Reference and scope
Further reading: Financial fact-level OCR evaluation: FinCriticalED. External documentation provides background for the topic; it does not establish Parzable’s implementation or performance. References checked 18 October 2026.
Examples and review decisions on this page are authored illustrations. Confirm unresolved business, payment, or accounting decisions through the responsible reviewer.
Put the workflow to work
Try your own invoice in Parzable
Create an account, define a schema, and test a small sample. Review extracted values before using them. See current plans and page allowances.
Try Parzable