A hard drive, three boxes and a spreadsheet
The client has sent everything they could find: a hard drive of emails, three boxes of paper, some photos of documents on a phone and a shared folder of contracts and drafts. The litigation team needs to review them and produce a list. A paralegal opens a spreadsheet and starts: date, document type, author, recipient, short description. Row by row.
Half way through, they realise many of the emails appear several times because they were forwarded and replied to, and some documents exist as a Word draft, a PDF and a scan of the signed copy.
Why listing takes so long
Most of the time is spent on mechanical work before any legal judgement: converting formats, reading dates off documents, working out who wrote what, and spotting duplicates. Large litigation uses specialist review platforms, but many small and mid-sized disputes are handled in spreadsheets, because the platform feels like too much for the case.
Scanned paper is the slowest part. Dates and authors have to be read by eye, and handwritten notes often have no clear date at all.
| Mechanical task | Why it is slow by hand |
|---|---|
| Dating documents | Dates written in different formats, or only in the content |
| Identifying author and recipient | Buried in letterheads, signatures or email headers |
| De-duplicating | Same document in several formats and forwards |
| Threading emails | Replies and forwards scattered through the export |
| Describing each document | Short description typed for every row |
What the manual list costs
Paralegal and trainee time goes into typing rather than reviewing. Deadlines for disclosure put pressure on the team, and mistakes creep in: wrong dates, duplicated entries, missed documents. Where the work is not fully recoverable, the firm absorbs the hours.
Fee earners also lose time checking a list they did not build and do not fully trust.
How we build a disclosure processing step
- Ingest: emails (PST or MSG files), scanned paper, photos and folders are loaded into one working set, with the original of each kept unchanged.
- Text and metadata: text is extracted, with recognition for scans and photos. Email metadata is read directly; for other documents, a language model proposes the date, author, recipient and document type, quoting where it found each one.
- Duplicates and threads: exact and near-duplicate documents are grouped, and email threads are rebuilt so the team can review a conversation once.
- Draft descriptions: each document gets a short neutral description for the list, which reviewers edit.
- Review screen: fee earners and paralegals mark each document for relevance, privilege and any notes, using categories your team sets. Nothing is marked automatically.
- Draft list: the tool exports a draft list in your required format from the reviewed set, with the working references kept for your records.
For very large reviews a specialist platform may be the right answer, and we will say so. This is aimed at the many cases that are too big for a spreadsheet and too small for a platform.
What the review feels like afterwards
The paralegal starts with documents already dated, grouped and described, so their time goes on checking and on flagging what matters. Duplicates are handled once. Fee earners review threads rather than individual emails. The list is produced from the reviewed data instead of typed separately.
The mechanical errors that used to appear in lists (wrong dates, duplicated rows) mostly go away, because they are no longer typed.
Signs your team would use this
- Disclosure lists are built in a spreadsheet by hand.
- Clients send documents as a mix of emails, scans and photos.
- Duplicates and forwarded emails swell the document count.
- Paralegals spend days typing dates and descriptions.
- Specialist review platforms feel too much for your typical case.