The problem
Discovery arrives as a pile of mixed files. Reviewers open each one to decide privilege and relevance, then retype the facts they will need later. The coding and the source file drift apart.
How the workflow runs
Classify the document
The classify step labels each file privileged, relevant, irrelevant, or other so the pile has an order before anyone reads for facts.
Extract the evidence
An extract agent pulls the key facts and references present in the document onto the run.
Leave privilege to counsel
The starter codes and extracts. Add a human review step on the canvas when counsel must confirm privilege and relevance before production.
What the hosted app shows
The discovery desk collects a packet and matter details, then shows the classification and extracted facts together. Counsel uses that page for the confirmation you add on the canvas.
Who it is for
Litigation teams and legal operations that need a first-pass coding of discovery before counsel reviews production.
Questions
It proposes a label and extracts facts. Privilege calls belong to counsel. Add a review step so that confirmation is part of the run.
Yes. Edit them on the classify step. The starter uses privileged, relevant, irrelevant, and other.
Contract review extracts terms from an agreement and runs a compliance check. This starter sorts a discovery set and extracts evidence.