Situation

Our client was assisting a plaintiff group in a class action filed in the Federal Court of Australia. The matter involved legacy employment records with potential evidence dating back over 50 years. The legacy documents had been digitised but traditional methods (like standard optical character recognition) were unable to make the documents searchable. Our client was going to have to manually review all the documents. Further, without any extracted text they would be unable to leverage any analytics features to make their review more accurate or more efficient.

Solution

Our class action experts identified a past custom solution we had developed for another matter, called Photo Search, which could assist. Our Photo Search tool was able to analyse the scanned images of the legacy documents and identify if handwriting was present. Our tool then was developed further to be able to extract searchable text for the handwritten annotations to an accuracy level of 75-90%. Traditional methods would have delivered a search accuracy of less than 20%. Once the process had been run over the document set, we were able to able to build the analytics indexes, run searches and otherwise support the team with review workflows that would otherwise have been unavailable to them.

Impact

The work our team did in applying the Photo Search solution allowed our client to:

• Cut down review time in being able to search documents and use analytics features

• Reduce the costs to their client as they relate to document review

• Decrease the time with which they can advise their client and respond to the Court/other parties in respect of the review process and their discovery obligations