What is QC review in eDiscovery?

QC review - short for quality control review - is a second look at documents that have already been coded, done to confirm that the calls made in first-pass review are consistent and defensible. It is a standard part of document review rather than a sign that something has gone wrong.

What a QC pass usually looks at

• Random samples drawn from each reviewer's batches, which produce an error rate you can point to if the coding is ever challenged.

• Targeted checks on the decisions with the most downstream consequence: privilege calls, documents marked non-responsive, and anything coded inconsistently across a family or a near-duplicate group.

• Reviewer consistency, by comparing how different people coded similar documents and correcting the coding manual where the disagreement is really an ambiguity in the instructions.

How teams measure it

The most common measure is the overturn rate - the share of sampled documents where the QC reviewer changes the original call. A rising overturn rate in one reviewer's batches usually points to a training gap. A rising rate across the whole team usually points to a coding manual that needs tightening.

QC works best running alongside review rather than at the end. Sampling the first few hundred documents in a batch catches a misunderstood issue code early, when re-reviewing is still cheap.

Where AI handles first-pass review, the same principle applies: sample the model's calls, measure how often a lawyer disagrees, and document the result. Related reading: The TAR 1 Reference Model: What Sedona Conference's New Framework Means for GenAI Review covers how validation frameworks approach measuring review quality.

Claira is an AI eDiscovery platform for document review at scale - you can book a demo to see how it fits alongside a QC process like this.