Technology-assisted review, also referred to as predictive coding in document review related to litigation, investigation, and regulatory compliance matters, has achieved judicial acceptance and has been empirically proven to make review both more efficient and more accurate than linear eyes-on review—achieving efficiency gains of 2:1, meaning that the team will look at only two documents to find each responsive document. Despite the clear benefits, only 25.9% of legal teams use TAR regularly. Similarly, despite budget and time constraints, 76% of legal professionals manually review documents in investigations.
Why is there a lag in adoption? Here we explore common barriers to TAR adoption and then discuss how to overcome these barriers for a more efficient and defensible TAR process.
TAR is still viewed as “black box”
Perhaps the greatest impediment to TAR adoption is its perception as “black box” technology. Historically, TAR engines have offered minimal transparency around why the TAR engine surfaced a particular document, which can lead to questions about the defensibility of the TAR review. Has the TAR engine really uncovered the relevant documents? What pockets of the unknown have been overlooked? How are results validated? These questions and others—when neither validated nor transparent to the legal team—can hinder adoption.
Alphabet soup
SAL, SPL, TAR 2.0 based on continuous active learning? Various TAR systems and approaches can make it difficult to understand the differences in approaches, and which protocol to use under which document review circumstances.
Eyes-on review of lengthy and complex documents
Even the most advanced TAR protocols based on continuous active learning require eyes-on review for prioritized documents surfaced by the engine. When those documents are lengthy and complex, the efficiency gains of using any TAR system can diminish with the time and effort by reviewers spent reading these documents in their entirety to pinpoint the portion that algorithm has identified as likely relevant.
3 ways to accelerate use of TAR
With the right process, people and technology, even the most reticent legal teams can often more effectively and efficiently meet their documents review goals using TAR with the optimal mix of people, process and technology.
- Leverage TAR experts
The devil is often in the details. TAR experts—including machine learning and statistics expertise often not resident in-house or at law firms–can provide support beginning-to-end. From developing and advising on the most efficient workflow to metrics and evaluation techniques to ensure high accuracy and defensibility, the right TAR advice can substantially reduce time and cost and improve accuracy. Where required, experts also can explain the technology and processes before judges or adjudicators.
- Optimize TAR processes to match your goals
To maximize the value of TAR to meet one’s document review objectives, it’s important to apply the appropriate methodology and approach, along with a well-thought-out workflow, to the circumstances. TAR experts can help legal teams understand the options that will help meet the legal team’s objectives: Is there time for the review team to lay eyes on every document that is going to be produced? Is the cost and effort of a certain review protocol proportionate to the value and importance of the matter? What are the risks?
For example, when responding to a third-party subpoena under a tight deadline, particularly where there is not a significant concern about sensitive, confidential or privileged information within the data set, the risk tolerance is often higher while budget thresholds are lower. In those circumstances, a full TAR review involving eyes-on every document being produced may not be optimal. A rapid analytic investigative review or a cut point review that leverage advanced analytics and rely on bulk coding and sample testing may significantly increase efficiency while keeping risk with acceptable parameters.
- Improve efficiency and transparency
TAR based on continuous active learning can improve efficiency by refining the accuracy of the algorithm with each coding decision made. Advanced companion algorithms such as contextual diversity explore all pockets of the document collection to ensure that no critical documents are overlooked–not just the documents that are conceptually closest to those already identified by the team as relevant. And, interactive and easy to interpret TAR dashboards provide TAR transparency with accurate and intuitive real-time reporting on the progress of the review, including recall and precision metrics, so a team knows when it can stop the review (and save further time and money).
The latest advance offers upfront summaries of the best passage in the document (versus just highlights of the search terms) to shave even more time off document review and coding and provide unprecedented TAR transparency.
Conclusion
If your legal team is still experiencing TAR FUD that is hurting your productivity and your bottom line OpenText OnDemand eDiscovery platforms and supporting services may be just what you need to take the next step in modernizing your litigation and investigation document review processes.
Learn more about eDiscovery solutions and Legal Tech & eDiscovery Services.
