Case study

How a Construction Disputes Team Turned 5,100+ Files into a Hyperlinked List of Documents in 2 Man-Days with BriefTech

A construction disputes team at a top Singapore law firm had two weeks, over the festive holidays, to prepare a List of Documents (LOD) for disclosure. The matter grew from about 2,000 files at appointment to more than 5,100 (~26,000 pages): scanned PDFs, Outlook emails, site photos, spreadsheets and CAD drawings.

BriefTech made 2,780+ scanned files searchable, identified 1,890+ engineering drawings, and populated dates and descriptions for the whole matter within a day. When the description format changed midway, the team hand-edited a sample of entries and BriefTech’s AI Refiner carried the new format across the rest of the matter.

The 4,800+ document LOD went out hyperlinked to source in 2 man-days, against an estimated 20.

Matter profile

Practice area: Construction disputes
Jurisdiction: Singapore
Team: A team of legal secretaries
Scope: List of Documents for disclosure: dates, descriptions, categories and source links
Matter size: 5,100+ documents, ~26,000 pages, ~30 GB
Constraints: Two-week deadline over the festive holidays, with staff on leave. Cost-sensitive client.

Key outcomes

2 man-days against an estimated 20

Indexing by hand runs at 200 to 300 documents a day, so a 5,000-document LOD takes about 20 man-days of secretary or junior associate time. BriefTech populated dates and descriptions for the whole matter within a day, and the team finished the reviewed LOD in 2 man-days.

Under half the cost for a cost-conscious client

The team produced the same work product for a low four-figure sum — under half what a junior associate spending a month on the index would have cost, a saving it could pass on to a client that watched every line of the bill.

A midway format change without the retyping

When the required description format changed midway, the team hand-edited a sample of entries to the new convention. BriefTech’s AI Refiner used them as examples and carried the format across the rest of the matter, so no one had to retype thousands of rows.

Deadline met without holiday overtime

The work landed over the festive holidays, with staff on leave. BriefTech ran overnight, unattended, so the team met the deadline without late nights or weekend shifts.

What the team weighed first

Before BriefTech, the realistic options were all poor. The team could index the 5,100 files by hand and lose most of the two weeks — and the holidays — to it. The firm’s existing document platform could store and hold the files but could not read a drawing title block or populate an LOD. Off-the-shelf OCR would make the scans searchable but still leave drawing numbers and dates — which sit in title blocks, not file names — to be transcribed by hand. None of these closed the gap between a two-week deadline and ~20 man-days of work, which is what sent the team to a workflow built for the deliverable rather than for storage.

How BriefTech helped

Challenges and solutions

Disclosure in a construction dispute means thousands of mixed-format files: drawings, site reports, correspondence and claims, many of them scanned. The supervising senior associate also had exacting rules for how each entry was described. For this matter, the team faced six specific challenges.

CHALLENGE 1

Short notice over the festive holidays

CHALLENGE: The team was asked to produce the LOD, with a date and description for every document, within two weeks. The request landed over the festive holiday season, with staff on leave and the remaining team short-handed.

SOLUTION: One administrator uploaded the files in bulk and BriefTech ran the workflow overnight, unattended: text recognition, date extraction, descriptions and categorisation. The populated index was ready for review within a day, leaving the rest of the time for checking and the client’s changes.

CHALLENGE 2

A matter that kept growing

CHALLENGE: The team was appointed to index about 2,000 documents. The client kept sending files through December and January until the matter passed 5,100. Keeping an index in step with a moving set of files across shared drives and local spreadsheets risks stale entries and missed documents.

SOLUTION: BriefTech held every file in one cloud workspace under a control number, so the index updated as files were added, removed or replaced. Only the 4,800+ documents marked for disclosure went into the final export.

CHALLENGE 3

Scans, emails, photos and CAD files

CHALLENGE: The matter mixed 4,400+ PDFs, 415 Outlook emails, 150+ site photos, 70+ spreadsheets and survey files, 50+ CAD and BIM files, Word documents and ZIP archives. More than 2,700 files were scans with no searchable text.

SOLUTION: On upload, BriefTech checked every file for searchable text and ran OCR on the 2,780+ that had none, leaving the 1,480+ already-searchable files untouched. It then sorted the matter by document type: 1,890+ engineering drawings, 1,340+ site, progress and technical reports, and 530+ items of correspondence.

CHALLENGE 4

Drawing numbers and dates

CHALLENGE: Drawings were the largest single group in the matter. Each had to be listed under its drawing number and the right date, which sit in the title block rather than the file name, and a drawing often carries several dates.

SOLUTION: A dedicated drawing step read the title block of each of the 1,890+ drawings and took the drawing number and date for its LOD entry, so no one had to open each drawing and transcribe it.

CHALLENGE 5

An exacting house style, changed midway

CHALLENGE: The supervising senior associate wanted every entry in a fixed format: emails by date, 24-hour time, sender and recipient company and a short description; drawings under their drawing numbers. After most descriptions had been generated, the required format changed. Retyping thousands of rows by hand would have consumed most of the remaining deadline.

SOLUTION: The team hand-edited a sample of entries to the new convention. BriefTech’s AI Refiner used those edits as worked examples and carried the format across the rest of the matter. The team’s final review and touch-ups fitted inside the 2 man-days.

CHALLENGE 6

A disclosure-ready index, linked to source

CHALLENGE: The LOD had to list every disclosed document with its number, date, description and category. Cross-checking a hand-built spreadsheet against folders of PDFs invites numbering errors and broken file paths.

SOLUTION: BriefTech exported the 4,800+ disclosed documents as an Excel LOD grouped by category, each entry numbered under the matter’s prefix and hyperlinked to its PDF, together with a ZIP of the PDFs. The export ran in about 20 minutes, unattended.

Who did what

Code handles the mechanical steps the same way every run. AI proposes the index and the sorting. The team sets the house style and decides what is disclosed.

Code (same result every run)

Bulk upload held under one matter control number
Searchable-text detection, with OCR only on the files that need it
The unattended overnight run and the hyperlinked Excel LOD export

AI proposes, lawyer decides

A date and description for every document
Drawing numbers and dates read from the title block
Document-type sorting: drawings, reports, correspondence
The new description format carried across the matter from a sample of the team’s edits

Lawyer only

The house-style description convention, and its change midway
Which documents go into the disclosed List of Documents
The final review and sign-off

Key stats (full matter)

Documents processed ~26,000 pages, ~30 GB: 5,100+
Documents disclosed in the LOD: 4,800+
Scanned files made searchable on upload 1,480+ already-searchable files left as-is: 2,780+
File formats PDF, Outlook email, images, spreadsheets, CAD/BIM, Word, ZIP: 7
Engineering drawings identified Drawing number and date read from the title block: 1,890+
Site, progress and technical reports: 1,340+
Correspondence and emails: 530+
LOD export Excel index and hyperlinked PDFs, unattended: ~20 min
BriefTech effort Index populated within a day: 2 man-days
Manual equivalent (estimate): ~20 man-days

BriefTech vs manual review

BriefTech vs manual indexing (this matter)

BriefTech vs manual indexing (this matter)
Measure	BriefTech	Manual
Effort Manual pace: 200–300 documents a day.	2 man-days	~20 man-days
First full draft of the index	Within a day	About four working weeks for one person
Cost Manual cost is junior associate time.	A low four-figure cost	Roughly double (junior-associate time)
LOD export Numbering and hyperlinking 4,800+ entries.	~20 min	By hand

Effort

Manual pace: 200–300 documents a day.

BriefTech: 2 man-days
Manual: ~20 man-days

First full draft of the index

BriefTech: Within a day
Manual: About four working weeks for one person

Cost

Manual cost is junior associate time.

BriefTech: A low four-figure cost
Manual: Roughly double (junior-associate time)

LOD export

Numbering and hyperlinking 4,800+ entries.

BriefTech: ~20 min
Manual: By hand

Manual figures are the team’s own estimates for indexing this matter by hand.

The impact

Rather than a month of transcribing drawing numbers and email subject lines into Excel over the festive holidays, the team had a populated index within a day and a reviewed, hyperlinked LOD in 2 man-days.

The midway format change, which in a manual workflow would have meant retyping thousands of rows, needed only a sample of hand edits. BriefTech’s AI Refiner carried the senior associate’s convention through the rest of the index.

For a client that watched every line of the bill, preparation cost came in at under half the manual estimate.

All documents were stored in Singapore, BriefTech’s default region, in isolated environments, with no data used for model training.

Get results like these on your next matter

Cut disclosure preparation from weeks to days: OCR’d, indexed, and hyperlinked to source, without the overtime.

Start free — 50 credits Schedule a live demo