6 December 2011 — This scenario unfolds every day: your company receives a request for production of documents. There are millions of files that are potentially relevant. You work with IT and your eDiscovery counsel to identify, collect and defensibly filter the data set down to exclude irrelevant documents. You use technology assisted review tools, e.g. clustering or predictive coding, to find the documents that are most likely to be relevant. Now you have your data set down to 250,000 documents and you are ready to produce them, right?
Probably not. Certainly you need someone to review a statistical sample of the documents for defensibility purposes. You probably also want humans to review at least some of the documents for privilege, trade secrets or to categorize documents in preparation for depositions or trial. Maybe you just want to know what you are producing before your opponent does, after all – the documents may all be relevant but some are more relevant than others.
For more click here.