Skip to main content
Note: Every Review Run is configuration-specific. The tags, models, instructions, and output fields shown in this documentation are examples only. Configure your Review Run to reflect your organization’s review objectives, criteria, and workflow requirements.
Before creating a Review Run, you must configure the review framework that will be used to analyze documents. Review Run configuration should be based on the review goals and criteria that define:
  • Key topics or themes
  • Categories or issues to identify
  • Relevant individuals, entities, or concepts
  • Matter- or project-specific risk factors
  • What constitutes helpful, harmful, or neutral information
Review Runs use this framework to evaluate documents consistently across the collection.

​
Review Run setup workflow

Complete the following steps before creating a Review Run:
  1. Define and approve your review framework.
  2. Create any required tag groups and tags.
  3. Create the model or models that will be used in the Review Run.
  4. Configure a Review Run using one or more models.
  5. Validate the Review Run configuration before processing documents.

​
Review Run components

Tags. Tags provide structured classification results that can be filtered, searched, and reported on. Text Classification models. Text Classification models assign tags or categories to documents based on the configured review framework. Text Generation models. Text Generation models produce written assessments, summaries, or other narrative outputs based on document content and review results. Review Runs. Review Runs connect one or more configured models to the document review workflow. Depending on your use case, a Review Run can use text classification models, text generation models, or both.