bit-x-bit Uses Analytics to Assist Law Firm’s Rapid Response to Government Subpoena
Technology Assisted Review (TAR) is a method of review that uses algorithms and artificial intelligence to identify potentially relevant documents based on reviewer input. As a reviewer identifies relevant materials during the review, the system uses that information to find other potentially relevant materials, prioritizing them for review and providing the attorney the most relevant documents up front. If you’re wondering how this works in practice, read on to see how bit-x-bit recently assisted a client in using TAR with great success.
When an energy-related technology company was subpoenaed by the Department of Justice regarding certain overseas transactions, the company’s law firm turned to bit-x-bit for the latest analytics technology to enable a defensible, rapid response with a minimum of attorney review time, using proven technology that would satisfy the DOJ’s lawyers.
bit-x-bit supplied and supervised Catalyst’s “Continuous Active Learning” (“TAR 2.0”) analytics technology, Catalyst “Predict.”® The ESI document collection of approximately 34,000 documents was the result of studious key word searching, but still only 10% of the documents were determined to be relevant and responsive (benchmarked by a random sample). Absent the use of analytics, the reviewers thus would have had to review ten documents to find one that was relevant – a grossly inefficient use of attorney time and talent, and a waste of time and money.
As the document review started and progressed, the “Continuous Active Learning” analytics noted the attorney reviewers’ relevance decisions and then continuously identified more and more relevant documents. After reviewing only 550 documents, the percentage of relevant documents presented to the reviewers increased from 10.9% to 43.6%, and then to 83.5% after reviewing only 2,144 documents, as reflected in the following chart:
Thus, the attorney reviewers were able to spend valuable time reviewing mostly relevant documents early in the review. Further, the analytics presented the relevant documents first, so that the review team did not even need to review 73% of the collection. This saved the client tens of thousands of dollars considering that the non-reviewed documents would have taken 140 hours to review. The documents which were not reviewed were statistically sampled to confirm that relevant documents were not “left behind” and served to demonstrate to the DOJ that the use of TAR 2.0 for review was reasonable and defensible, and complied with DOJ standards.
The savings on this review were substantial. Not only that, the client was able to get to the most relevant documents early in the review process and complete the review in less than half the time a linear – or traditional document-by-document – review would have taken.
We would love to do the same for you. Call us now to talk about how we can help your eDiscovery needs with TAR!