Understanding Continuous Active Learning (CAL) in Modern Document Review
One of the most impactful AI-driven tools in today’s review platforms is Continuous Active Learning (CAL). CAL helps streamline review by constantly learning from the documents you’ve already tagged and using that information to prioritize what you should look at next.
At its core, CAL evaluates the documents you’ve marked, whether Responsive, Non‑Responsive, or even issue-specific tags, and then scores the remaining documents on a scale from 0 to 100. That score reflects how likely each document is to match the pattern of what you’ve already identified.
This makes CAL useful in two powerful ways:
- Positive/negative reinforcement: As you tag documents Responsive or Non‑Responsive, CAL quickly adapts and pushes the most likely Responsive documents to the top of your queue.
- Issue‑based discovery: CAL isn’t limited to binary tags. You can run it on single‑issue tags to surface documents similar to your “hot docs,” helping you expand an issue set with far less manual digging.
The result is a smarter, faster, and more targeted review process—one that continuously improves as you work.