Storybuilder and Writing Assistant (Everlaw AI Assistant Beta)

This article is a deep dive on automated descriptions and Writing Assistant, which are features available in the Everlaw AI beta. 

  • If you're looking for an overview/starter guide for all beta features, see this article.
  • If you're looking for a general overview of the beta program itself (including how to join), or information about our privacy and security standards, please see our Beta FAQ
  • If you have feedback or questions, feel free to email us at We love hearing from our users.

Table of content


🤖 AI descriptions

Storybuilder document entries on the Timeline have a description field. If:

  1. This description field is not already populated with a human-written description, and
  2. An AI summary/description is available for the document

then, Everlaw will display the AI description in this field. This is the same description that is generated from the Everlaw AI summary task, which you can learn more about here

Screenshot 2024-05-02 at 11.42.38 PM.png

AI descriptions will be displayed in italicized font and have a sparkle icon indicator. You can freely edit the description field. Keep in mind that if you "edit" an AI description from the Timeline, you are not changing the underlying AI description itself. Instead, this action is equivalent to:

  • Making a copy of the AI description
  • Applying your edits to it
  • Adding the finalized version into the field as human-generated text

✏️ Writing Assistant

Writing Assistant is a flexible tool that can help you synthesize, structure, and brainstorm against the evidence curated in Storybuilder. 

Accessing the Writing Assistant

You can access the Writing Assistant by clicking the “Assistant” button in the toolbar of any Draft or Deposition. This will open a side panel where you can create new Writing Assistant tasks and view past tasks.

Everlaw preserves and displays a history of all Writing Assistant tasks for each Draft or Deposition, including (1) the prompt and configuration settings used to create the task and (2) the full result from that task. 

Creating new tasks

You can create new tasks by clicking the “Compose” button in the side panel. This will open a flow with two steps. 

Step 1: Evidence selection

Select which documents you want to use as part of the task. Everlaw supports up to 500 documents per task. There are three options:

  • All documents in the Story
  • All documents currently added to the draft or depo
  • All documents labeled with a given label in the Story

Note 1: The full text of the documents included in a task might not be used in the task. For example, if hundreds of documents are included in the task, each document will be represented by short descriptions of their content rather than their full text. Alternatively, if the task includes a handful of shorter length documents, longer summaries or full document text will be used to represent the content of the documents.

Note 2: The resulting output will not include references to all documents included in your selection, unless that set is very small (< 10 documents). The primary use case for making an expansive selection at this step is to allow Everlaw AI to pick-and-choose the documents it believes are most relevant to your request. If you already know you want to analyze a particular set of documents, we encourage you to make a narrower selection at this step. 

Step 2: Task configuration

In this step, you will configure options that determine both the form and substance of the task. The table below summarizes the available options.

Form Description Configuration options
Memo The resulting generation will be a longform piece of writing with in-line citations. The memo comprises multiple sections, each of which has multiple paragraphs. 

Tone: The tone of the memo or outline, which can be either analytical/dispassionate (“analyzes”) or persuasive (“argues”). 

Substance: You must also describe what you want the memo/outline to analyze or argue, along with any additional details or instructions. 

Outline The resulting generation will be an outline, with concise descriptions of what should be covered in each of the sections and subsections, and citations to relevant pieces of evidence. 
List The resulting generation will be a bullet-point list (with potential sub-bullets, depending on the configuration) with relevant citations for each item in the list. 

Each item is: You must describe what each item on the list should be or represent (for example, a person, a question, an event, an example of something, etc.)

Item details: You can optionally specify what additional details or analysis you want included for each item. These will be added as sub-bullets. 


Sort order: You can optionally specify if you want the items ordered in chronological or reverse chronological order. 

Rewriting, or writing against, existing content

If you've highlighted content in the draft body, the “Rewrite” option will become available in the Writing Assistant. Clicking “Rewrite” will open a configuration page where you can provide instructions on how to rewrite, or write against, the highlighted text. You can also select additional pieces of evidence you want included as part of the task (evidence found in the highlighted text will be included automatically). 

 This is a flexible option that can be used in a variety of ways. For example:

  • You can highlight a paragraph and ask for Writing Assistant to come up with 6 different variations that differ in terms of style and formality
  • You can highlight a section of an outline and ask Writing Assistant to flesh it out into a full draft, while taking into account additional pieces of evidence
  • You can upload an opposing party’s brief into the draft body, highlight some paragraphs where a particular argument is developed, and ask the Writing Assistant to come up with some counter arguments taking into account Story evidence
  • You can highlight text that represents a specific production request item and ask the Writing Assistant to list documents in the story that are relevant to that request

In-progress generations and viewing output

Once you click “Generate” or “Rewrite”, the task will be sent to the generation queue and you'll be brought to the output page for the task. The top half of the output page shows all the configuration details for the task and the bottom half displays the resulting generation. In-progress generations will be streamed word-by-word into this area.

Because Writing Assistant tasks are completed in the background, leaving this view or closing the draft/Everlaw will not affect the progress of the task. From this view, you can:

  1. Start a new task using the same task configurations.
  2. Insert the completed output into the draft at the current cursor position or at the end of the draft body.
  3. Copy the contents of the output to your clipboard.
  4. Expand the output container to see more of the generation.
  5. Delete the task.

Task history

Each task generated within a draft or depo is represented as an entry on the task table. Each entry has identifying information, including (1) the user instructions, (2) the datetime of generation, (3) the user who created the task, and (4) the current status of the task (completed, generating, failure). Clicking a task entry will open the full output and configuration view for that task, described above. 

You can sort the task history by date and filter by user.


Have more questions? Submit a request


Article is closed for comments.