Duplicate documents are not just messy.
They create audit risk:
- teams approve the wrong version
- payments get supported by the wrong attachment
- evidence becomes ambiguous (“which one is correct?”)
Start here:
The simplest reliable dedupe signal: checksums
Filenames are not reliable.
Two different files can share a name, and the same file can be uploaded with a different name.
A checksum (like SHA-256) lets you detect duplicates deterministically.
What to do when you detect duplicates
Do not delete immediately.
Instead:
- Confirm whether the duplicate is identical (checksum match)
- Decide whether to keep one canonical record (and tag it)
- Maintain version history if the “duplicate” is actually a revision
Next steps
If your evidence library is drifting, fix structure first: