Everlaw

Storybuilder is Everlaw’s narrative-building workspace for organizing the evidence, facts, people, deposition material, and written work product that make up a case. It lets legal teams build fact timelines, organize evidence, prepare depositions, analyze testimony, and develop drafts while keeping these materials connected to the underlying case record. Teams can also use transcript search, deposition designations, exhibit lists, and collaborative drafts.

This feature generates first drafts, evidence outlines, legal memos, comparison tables, and other case-preparation materials from document sets selected by the legal team. It can also analyze deposition transcripts and create question outlines or summaries with page-line citations that point back to the underlying testimony.

Everlaw Predictive Coding uses machine learning to learn from a legal team’s existing document-coding decisions and predict which remaining documents are likely to be relevant. It prioritizes high-value evidence while allowing teams to de-prioritize potentially non-responsive files, helping reviewers focus their efforts on the most important materials first. Teams can also reuse previously trained models across matters, overlay predictive coding results with Everlaw’s clustering visualization, and validate model performance using built-in precision and recall metrics.

Review Assistant provides document-level analysis inside Everlaw’s review workflow, including summaries, topic analysis, custom extractions, sentiment analysis, and natural-language Q&A. Custom extractions can pull specific information such as people, organizations, dates, numbers, or other user-defined data points from documents, while document Q&A returns answers with citations to the relevant source passages.

This feature automatically organizes documents by conceptual similarity, rather than requiring reviewers to define keywords or categories first. Everlaw represents these groups visually, allowing users to move from a high-level view of the document population into individual clusters, related documents, and specific evidence. The visualization can handle up to 25 million documents on a single screen and can be combined with search results, coding decisions, and Predictive Coding scores.