Comprehensive AI-Informed Review

Deep Dive for Early Insights

• Use Deep Dive to ask big-picture questions about your data set and quickly identify key themes, issues, and documents to prioritize for review.

• Start by querying all your documents, or you can narrow your data set using Search, Data Visualizer, Search Term Reports, or any platform tool that returns a set of documents.

• Ask open ended questions about the data set, for example:

  • Why did Acme Corp collapse?

  • What are the main categories of alleged misconduct?

  • Who are the key players in this matter?

  • When did Alex Carter approve the misuse of funds?

• Use the documents surfaced by Deep Dive to inform your review set, adding them to Assignments or Binders for human review.

• Learn more about using Deep Dive:

Streamline review with Everlaw AI Assistant

Use Review Assistant during document review to generate summaries, descriptions, and document-specific answers that help reviewers triage and understand documents faster.

• In the review window, click the sparkle icon to open Review Assistant

  • Summary & topics: Use summary & topics to quickly understand what a document is about by generating a short overview and highlighting the main themes, entities, and sentiment—for example, reviewing a long document to identify the key issues before you code the document.

  • Custom extraction: Use custom extraction to pull specific information from documents into a structured format when you know exactly what details you want to capture—for example, extracting drug names, dosages, and providers from patient records.

  • Document Q&A: Use document Q&A to ask a targeted question about a single document and get a direct answer based on the document text—for example, asking a lengthy policy manual whether it mentions vendor contract risk mitigation and where that appears in the document.

• In the results table, add descriptions or custom extractions as a column for quick triage.

• If you have AI credits, you can batch generate summaries and custom extractions from the results table.

Set up Coding Suggestions

• Project admins will need to set up coding suggestions the Project Settings, under the Everlaw AI tab on the left.

• Each coding category and code will need a written description of the code criteria in natural language. Think of this as the review protocol the AI will follow, including:

  • What the code is intended to capture

  • Concrete examples of in-scope and out-of-scope language

  • Any key players, dates, or terminology relevant to that code. 

• Pilot coding suggestions on a small sample so reviewers can inspect the suggestions, reasoning, and highlights before adjusting criteria and building confidence in the results.

• Read through this guide for more help on writing coding suggestion criteria:

Scale Coding Suggestions for Efficiency and QC

Scale coding suggestions across your review set to speed up primary review and focus quality control on documents where AI suggestions and human coding do not align. Note: this step will require AI credits.

• From a results table of your active review set, click the Everlaw AI button on the middle toolbar, select Coding Suggestions and configure a batch action.

• Use the output of the coding suggestions to accelerate primary review by quickly applying high confidence suggestions.

• “Actionable” suggestions indicate coding where AI and humans disagree: use the filter at the top of the coding suggestions column to quickly find documents to QC.

• Learn more about reviewing with coding suggestions:

Storybuilder and Writing Assistant

Move from review to case strategy by using Storybuilder and Writing Assistant to organize fact timelines, create a cast of of characters, generate case outlines or deposition questions grounded in selected documents.

• Before adding documents to Storybuilder, configure your AI summary settings:

  • In Project Settings, under the EverlawAI tab on the left, select Document Summaries and enable “Automatically summarize new Story objects”.

• In Deep Dive, directly copy responses with their document references by clicking the quill icon to add to an already-created Draft or Deposition

• In Drafts and Depositions, click the sparkle icon to open Writing Assistant to begin generating work product and analysis based on your prompts or pre-set templates.

• Learn more about Writing Assistant:

Working with Deposition Transcripts

Use natural language questions to query deposition transcripts to find key testimony, witness inconsistencies, and more.

• Create a deposition object in Storybuilder for each deposition and upload the transcript (or add a transcript to an already-created deposition).

• Navigate to the Summary tab of the deposition.

• Open Writing Assistant and find the Transcript analysis tab - select “prepare for analysis”

• Use the Transcript analysis option to generate summaries, analyze witness discrepancies within a single deposition, or ask custom questions such as “What did Smith assert regarding pricing?”

• On the Depositions page, use the “Ask across Transcripts” button to pose questions across all uploaded transcripts, such as “Find any testimony that corroborates or disputes Smith’s assertions on pricing”