As timelines and budgets continue to shrink and the volume of ESI in eDiscovery and investigation matters continues to grow exponentially, the substantial time and cost savings delivered by technology-assisted review (TAR) is more critical than ever. Put simply, there is too much data to rely solely on eyes-to document review an entire collection.
TAR is proven to increase review efficiency and reduce cost.
It has been empirically proven, and now widely accepted, that the use of TAR substantially reduces review costs. By employing an algorithm to surface the documents most likely to be responsive, teams can drastically reduce the number of documents that require eyes-on review and know confidently that the majority of the responsive documents have been located. Rather than being lost in a sea of data and having to blindly review the data document by document, TAR offers an intelligent way to prioritize and ultimately reduce the overall data requiring review.
TAR adoption rates are still lagging.
Surprisingly, eDiscoveryToday’s 2022 State of the Industry Report revealed that only 25.9% of respondents use predictive coding in all or most of their cases, while 36.3% of respondents use it in very few or none of their cases. Similarly, a recent Compliance Week survey found that despite budget cutbacks and time constraints, 76% of compliance professionals are still employing manual review of documents in ESI investigations.
So why, despite the quantifiable benefits of TAR, is there still lingering FUD around its use?
Why are most legal teams still not using TAR despite demonstrable benefits?
The greatest impediment to full and consistent adoption of TAR is persistent FUD around the TAR “black box.” People, especially legal professionals, are naturally skeptical of what they can’t see. Keywords are easy to spot (especially when they are highlighted) but understanding the complexity of why the TAR engine thinks that one document is conceptually similar to documents previously coded relevant has previously been a mystery. Historically, TAR engines have offered minimal transparency, leaving reviewers in the dark about why the TAR engine has surfaced a particular document. This leads to mistrust and discomfort when elusion tests suggest that review can be halted when the agreed upon percentage of the responsive documents have been located.
Offering transparency, greater efficiency and cost-savings in document review
OpenText™ Insight Predict, a TAR based on continuous active learning, now offers a solution to the TAR black box conundrum. Insight Predict best passage highlighting, a new and unique feature released in Cloud Edition (CE) 22.1, offers greater efficiency and transparency to help teams conduct document review with maximum efficiency, accuracy, and peace of mind.
The Insight Predict TAR engine removes FUD by highlighting the best passage in the document so that reviewers can immediately zero in on the contents of the document that the algorithm has identified as being similar to documents previously coded and relevant to the new research topic. Review and QC teams can make their own assessment of the TAR engine’s accuracy when they can see what passages the TAR engine identifies as potentially important. They also may feel more confident when the system indicates that the likelihood of locating any additional relevant documents in the remaining collection is too low to warrant further review.
Best passage highlighting delivers even greater TAR efficiency
While Insight Predict is exceptional at quickly and reliably surfacing relevant documents, until this breakthrough, there was room for even greater efficiency. Without best passage highlighting, reviewers need to skim the entirety of the document to determine responsiveness. With short email messages, it is relatively easy for reviewers to skim the document to determine whether they agree that the document surfaced by the TAR engine is truly relevant. But when the review involves a large volume of lengthy documents, reviewers may find themselves stuck on the document, needlessly burning through time and budget as they search the document trying to figure out why the TAR engine thinks it is relevant.
Best passage highlighting immediately draws the reviewer’s attention to the key passage enabling them to determine responsiveness faster than ever. The result is substantial time and cost savings.
With this revolutionary new enhancement, teams will now have even more reason to rely on Insight Predict TAR to improve efficiency, reduce costs, and make eDiscovery and investigation document review easier than ever.
Insight Predict – Not all TAR is created equal
The OpenText™ Insight Predict module within the Insight eDiscovery and investigation platform provides advanced TAR for productions, investigations, early case assessment, witness and issue preparation or privilege QC. Insight Predict is proven to lower the total cost of review (TCR) with an efficiency rate as high as 2:1 – meaning that for every two documents reviewed one is responsive. Insight Predict developed the industry’s first commercially available continuous active learning (CAL) algorithm which continuously learns what’s important to the matter and becomes smarter as review progresses. This technology also features a contextual diversity algorithm that eliminates the risk of missing relevant documents by searching pockets or clusters of contextually diverse data to ensure that potentially relevant documents are surfaced in the shortest time. Contextual diversity locates potentially relevant documents with demonstrably greater efficiency and reliability than random sampling.
To learn more about how Insight cloud-based eDiscovery review platform can help your team realize greater review accuracy, efficiency and peace of mind in your next document review consult one of our experts today.