As document volumes continue to grow across litigation and investigation matters, legal teams face a consistent operational challenge: reviewing large volumes of information accurately, efficiently, and in a manner that supports defensible outcomes while managing review timelines and costs. Technology-assisted review (TAR) is one of several tools organizations may utilize to help manage large document populations and prioritize review efforts. The effectiveness of TAR depends on the technology, the review workflow, validation procedures, quality control measures, and attorney oversight applied throughout the review process.
Key Takeaway
Technology-assisted review is most effective when incorporated into structured review workflows that combine attorney judgment, validation procedures, and quality control oversight. While TAR is widely used across litigation and investigations, defensibility depends on how the workflow is designed, documented, and supervised throughout the review process.
What Is Technology-Assisted Review?
Technology-assisted review, also referred to as predictive coding or computer-assisted review, is a process that uses machine learning algorithms to identify documents likely to be relevant within a larger dataset. The process begins with attorneys coding a seed set of documents as responsive or non-responsive to the discovery requests at issue. The software analyzes those coding decisions, identifies patterns, and develops a predictive model that can then be applied across the full document population.
The model categorizes documents by likelihood of relevance, and the review team uses those rankings to prioritize and structure the broader review. The degree of accuracy the model achieves is assessed through validation testing, which informs decisions about what additional human review is warranted before production.
How TAR Works in a Structured Review Workflow
Technology-assisted review is typically incorporated into broader document review workflows that include attorney oversight, validation procedures, and quality control measures. The structure of those workflows helps support both review accuracy and defensibility.
A well-structured TAR workflow typically includes the following components:
Seed Set Development
Attorneys review an initial document sample and apply consistent coding decisions that train the predictive model. The quality of the seed set has a direct effect on model performance, making attorney involvement at this stage a foundational element of the process.
Iterative Model Training and Validation
As the model develops, review teams assess its performance through sampling and validation protocols, refining coding decisions and retraining as needed. Documented validation testing is a significant factor in TAR defensibility.
Prioritized Review of Ranked Documents
The model’s output is used to structure the review, directing attorney attention toward the highest-ranked documents while establishing a documented basis for how lower-ranked materials are handled.
Quality Control Throughout the Workflow
Defined QC procedures applied throughout the workflow help support consistency in coding decisions and provide documentation of the review methodology. Structured quality control also allows review leaders to identify trends, address coding inconsistencies, and refine review protocols as the project progresses.
Defensibility Considerations in Technology-Assisted Review
Courts have addressed TAR defensibility in several significant decisions. In Da Silva Moore v. Publicis Groupe, the Southern District of New York recognized TAR as an acceptable discovery methodology when properly implemented and overseen. The Rio Tinto v. Vale decision reinforced judicial acceptance of TAR when supported by appropriate validation, transparency, and review protocols.
These decisions reflect an important principle: defensibility is a process question, not a technology question. A TAR workflow that is poorly designed, inadequately validated, or inconsistently applied introduces the same concerns that may arise with any review methodology lacking documentation and oversight. Conversely, a structured TAR process with defined protocols, attorney oversight, and documented quality control procedures provides a strong foundation for explaining and supporting review decisions.
The Sedona Conference has also published guidance addressing technology-assisted review and defensible discovery workflows, emphasizing the importance of transparency, validation, and documented review methodologies.
Legal Team Support
Baer Reed supports law firms and corporate legal departments with technology-assisted review services designed to integrate attorney oversight with scalable review workflows. Through structured TAR implementation, defined quality control procedures, and documented validation protocols, Baer Reed helps legal teams manage high-volume discovery in a manner that supports both operational efficiency and defensible outcomes.
Contact us to learn how our document review services support technology-assisted review workflows in litigation and investigation matters.
FAQs
TAR is most commonly used in matters involving large volumes of information where manual review of every document would be impractical within the available timeline or budget. It is particularly well-suited to matters with defined relevance criteria that allow for consistent coding decisions during seed set development.
Read More: Advantages of Integrating Legal Managed Services with eDiscovery
Technology-assisted review workflows rely on clearly defined review criteria, coding guidance, escalation procedures, and validation protocols. Establishing these standards before review begins helps support consistency across review teams and provides a framework for evaluating review outcomes throughout the project lifecycle.
Read More: How to Create an Effective Guidance Document for eDiscovery
Attorneys help develop and code the seed set that trains the predictive model, make escalation decisions on complex documents, oversee validation testing, and maintain quality control oversight throughout the review process. Attorney involvement remains a central component of a well-structured TAR workflow.
Read More: Modern Efficiency: Combining Attorney Teams and AI Support
Validation testing involves sampling documents from across the review population, including documents the model ranked as non-responsive, to assess the accuracy and consistency of the model’s predictions. The specific methodology may vary by matter, but the objective is to document how the review team assessed model performance throughout the workflow.
Read More: Quality Assurance in Legal AI: Validating Models, Preventing Drift
Quality control procedures often include reviewer calibration, targeted QC workflows, coding consistency checks, validation reviews, and supervisory oversight. These procedures help legal teams assess review outcomes and maintain consistency across large document populations.
Read More: Understanding the QC Process in Document Review
TAR and privilege review are often conducted as separate workflows within a larger review process. TAR may be used to identify responsive documents, while privilege determinations are addressed through separate review protocols designed to identify and protect privileged communications and work product. Organizations managing large review populations often establish distinct validation and quality control procedures for both workflows to support defensibility throughout the review process.
Read More: Creating Defensible Privilege Review Processes in eDiscovery
Mr. Reyes graduated with honors from the Ateneo de Manila University, where he received the Procter and Gamble Student Excellence Award. He obtained his Juris Doctor degree from the Ateneo de Manila School of Law. During law school, Mr. Reyes was part of the Philippine delegation to the Willem C. Vis International Commercial Arbitration Moot held in Vienna, Austria. He was also a member of the Ateneo Society of International Law and the St. Thomas More Debate Society. He completed his internship at the Public Attorney’s Office. He wrote a thesis entitled: “To Kill A White Elephant: An Analysis of the Fiduciary Exception to the Corporate Attorney-Client Privilege”. Mr. Reyes is admitted to practice law in the Philippines and the State of New York.
Matthew Hersh earned a B.A. in Political Science from Columbia University in 1990 and graduated cum laude from Georgetown University Law Center in 1999. He also holds a master’s degree in international relations from the Georgetown University School of Foreign Service.
Cap. Avi Levak (Res. IDF) graduated from from Israel’s prestigious Ben-Gurion University of the Negev with a Bachelor of Science in Computer Science and Mathematics. He is also a Leadership and Communication coach trained in TuT coaching by Alon gal in Israel. Avi specializes in high-level, in-depth analysis of business and client needs, within systems and software strategy and architecture.
Ms. Lardizabal-Manzano is a graduate of San Sebastian College-Recoletos, where she earned her B.A. in Political Science. In 2003, she received her law degree from Lyceum of the Philippines and was admitted to practice law in 2004.
Mr. De Guzman graduated from San Beda College with a degree of Bachelor of Arts Major in Economics and received his law degree from San Beda College of Law. He is multilingual and is fluent in three languages: Chinese, Filipino, and English. He was admitted to the Philippine Bar in 2003.
Ms. Aquino-Batallones obtained a Bachelor of Arts degree in Development Studies (with Minors in Global Politics and Hispanic Studies) from the Ateneo de Manila University. In 2011, she received her Juris Doctor degree from Ateneo de Manila University School of Law. During law school, she interned at Romulo Mabanta Buenaventura Sayoc & de los Angeles then became an intern of Ateneo Legal Services Center’s Clinical Legal Education Program.
Ms. Cruz-Anonuevo graduated cum laude and top nine in her batch from Miriam College with a degree of Bachelor of Arts in InternationalStudies. She obtained her Juris Doctor degree from Ateneo de Manila University School of Law in Rockwell. During law school, she interned in Rivera, Santos, Maranan & Associates. She was also part of Ateneo’s Labor Law Bar Operations. She wrote her thesis on, “Stealing Privacy: Limitations on Media’s Photographic Invasion.,” Ms. Cruz-Anonuevo is admitted to practice law in the Philippines.
Ms. Tyler graduated cum laude from Georgetown University and received her law degree, cum laude, from Georgetown University Law Center. During law school, she interned at the United Nations Economic Commission for Europe. She also worked on The Tax Lawyer journal and was a member of the award-winning Barristers’ Council Mock Trial Team. Ms. Tyler is admitted to practice law in the State of California and the District of Columbia.