Technology Assisted Review

Technology Assisted Review (TAR) is a process in which computer software—guided by expert reviewer input—electronically classifies documents to expedite the organisation and prioritisation of a document collection. The software’s classification may encompass broad issues such as discovery responsiveness, privilege, confidentiality, or other designated matters. TAR can significantly reduce the time and cost associated with reviewing electronically stored information (ESI) by limiting the amount of human review needed on documents that are classified as potentially non-material.

Review is the most expensive phase of eDiscovery, and technology has sharply cut its cost — recent industry data puts AI-assisted review at around $0.11–$0.50 per document, versus $1.50–$3.00 for manual review only a couple of years earlier (ComplexDiscovery, 2026). Cutting the volume a team must review is where TAR delivers the greatest savings.

However, the effectiveness and defensibility of TAR depend heavily on working with a knowledgeable vendor who understands the nuances of the software and the legal standards governing its use. An experienced TAR provider not only designs and implements a defensible workflow, but can also provide transparency about how the system was trained, validated, and monitored. This is critical in ensuring that the process will withstand scrutiny from opposing parties or the courts.

Vendors with a proven track record—including having supplied expert affidavits in prior cases—can help establish confidence in the process, both for the parties and the judiciary. Their expertise can be decisive in demonstrating that TAR was applied in a fair, proportionate, and technically sound manner. In practice, this can mean the difference between a streamlined, cost-effective discovery process and one vulnerable to disputes, delays, or judicial rejection.

Technology Assisted Review FAQs

What is the difference between TAR and Continuous Active Learning (CAL)?

TAR is the broad category of using machine learning to classify documents for review. Continuous Active Learning (CAL) is a specific, modern TAR protocol in which the model retrains continuously from reviewer decisions and constantly re-prioritises the remaining documents, rather than training once on a fixed sample.

When is TAR worth using?

TAR delivers the greatest benefit on larger document sets, where reviewing every document in a linear review would be slow and costly. By prioritising the documents most likely to be relevant, it reduces the volume a team needs to review while improving consistency.

How do you keep a TAR process defensible?

A defensible TAR workflow depends on an experienced provider who designs and validates the process and can document how the system was trained, tested and monitored — transparency that lets the approach withstand scrutiny from opposing parties or the court.

Considering TAR for a document review? Contact us or call 0800 WITNESS (0800 948 637).