Expected US Release: Jul 28, 2023
Expected International Release: Aug 4, 2023
Expected Federal Release: Aug 11, 2023
With this release, we’ve added a communication visualizer, video depositions, and more — read on for more information about the features coming out this month! If you would like to learn more about the features in this release, join us for a live training session.
User-facing features in this release:
- Communication Visualizer
- Video depositions
- Partial redaction of pre-defined PII searches
- Scheduled database suspensions
- Jump to row number in results table
- Jira connector improvements
- Predictive coding improvements
- API for directory updates in Legal Holds
- Code as previous project-level setting
- Allow empty prefix for PDF uploads without load file
- Improved error reporting for Microsoft 365 uploads
- Improved legal holds directory table
Finally, we're excited to announce the launch of Everlaw AI in closed beta. Everlaw AI is a suite of capabilities, powered by Generative AI, that can help you work more efficiently and effectively across review and writing/analysis. Our goal with the closed beta is to put this functionality in the hands of early users, and evolve and improve our offering based on feedback and learnings along the way. For more information about Everlaw AI, please see this website.
Communication Visualizer, Everlaw’s newest addition to Data Visualizer, is an interactive and filterable visualization of email communications. Communication Visualizer automatically analyzes the names, email addresses, and domains of emails in a given search to identify key players and their patterns of communication. You can see who was communicating with whom, with what frequency, and how those patterns evolved over time, before exploring the underlying emails.
Communication Visualizer is directly integrated with Everlaw’s Data Visualizer. You can visualize email data from the search page or from the results table. Communication Visualizer generates graphs that involve three main components:
- Nodes (circles) that represent a name, email address, or domain;
- Edges (lines with arrows) that represent the volume and proportional directionality of the emails exchanged between nodes;
- The “Others” node, the quarter circle in the top left corner of the visualization, represents all communications between nodes displayed in the visualization and other nodes in the document set.
To access Communication Visualizer from the search page, click “Open data visualizer” after building your initial search. From a results table, select the “Visualize” option from the top toolbar. Once in Everlaw’s Data Visualizer, click the “Communication” tab on the left side of the screen to generate a Communication Visualizer graph.
Learn more about Communication Visualizer here.
You can now upload and view synced deposition videos alongside their deposition transcripts in Storybuilder. Highlight a snippet of deposition testimony in transcripts to automatically create an exportable matching video clip.
After uploading a deposition video, you can view and play the video from the top right of the transcript view of a deposition. You can create clips of the video to add to a story, view highlights in the new highlights panel, and export video clips.
Learn more about video depositions here.
Partial redaction of pre-defined PII searches
You can now partially redact certain types of personally identifiable information (PII). For example, you can elect to redact only the username within an email address or everything but the last four digits in a numerical identifier.
You can create partial redactions from persistent and custom hit highlights, from the text selection tool, and in batch redactions.
Learn more about PII redactions here.
Scheduled database suspensions
You can now schedule a suspension for the end of the billing month. Projects in a database with a scheduled suspension will remain active until the end of the month UTC time, and will display a badge indicating their coming suspension.
To schedule a database suspension, navigate from the Overview tab of Database settings and click the “Suspend” or “Manage suspension” button to see various suspension options.
Learn more about database suspensions here.
Jump to row number in results table
You can now jump to a specific row in the results table by typing the row number into the selection box in the header and hitting Enter.
After the jump, the row will be automatically highlighted in the table. The row will appear at the bottom of the table if jumping from a lower row number, or at the top of the table if jumping from a higher row number.
Learn more about the results table here.
Jira connector improvements
When uploading data from Jira using Everlaw's Jira cloud connector, Everlaw will now indicate if your Atlassian account does not have access to issues in specific projects from the selected Jira site.
Learn more about the Jira cloud connector here.
Predictive coding improvements
A number of improvements have been made to Everlaw’s Predictive Coding! With this release, you can now:
- generate prediction scores after reviewing 200 qualified documents with at least 50 relevant and 50 irrelevant pursuant to the model’s criteria;
- track progress towards generating initial prediction scores and easily create training sets of documents that can be reviewed to meet the training threshold under the new “Prediction generation progress” section; and
- identify a model’s update status (e.g., Idle, Model queued for update, and Updating).
With this release, we’ve also made some notable visual changes and updates to copy related to Predictive Coding.
- Within the “Training” section on the predictive coding model page, the “ALL TRAINING DATA” section has been renamed to “Model Data” for clarity.
- Document counts for excluded documents and ineligible documents will now be displayed under a new section on the predictive coding model page titled “Excluded & Ineligible.”
- The existing sections for model performance, holdout set, and model history will now be located within the new “Performance” section (instead of the “Results” section).
- The button to toggle between Basic and Rigorous performance metrics types is now located in the top right corner of the new “Performance” section.
- The button to toggle between the graph view for historical performance and holdout size will now be found in the top right corner of the “History” section within the new “Performance” section.
- Legacy training sets (i.e. training sets created prior to the release of this feature) will appear across all models within the project and will be identified under the “Training sets” section with a red asterisk.
- “Performance statistics” has been renamed to “Performance metrics.”
Learn more about generating initial prediction scores for your predictive coding model here. For an introduction to predictive coding, see our beginner’s guide to predictive coding here.
API for directory updates in Legal Holds
You can now use the Everlaw API to manage the employee directories used in legal holds. Organizations can use this functionality to automate updates to their directories in Everlaw without creating a direct connection between Everlaw and their directory platform.
The new addition to the API allows you to automate a CSV upload into the Legal Holds system. You can import new people into the directory as well as overlay information onto existing people.
Whether the directory is uploaded automatically via the API or manually in the Legal Holds tab, the directory can include the following properties: name, email, business phone, mobile phone, start date, end date, role, company name, employee status, manager name, manager email, address, and notes.
Code as previous project-level setting
You can now set the behavior for Code as Previous at a project level. Project-level Code as Previous settings can be modified under the Codes tab of the project settings page. These settings mirror the user-level settings.
Whenever a project admin makes a change to these Code as Previous settings, all users on the project will have their Code as Previous settings immediately updated to conform. However, users can still change their individual settings, and the most recent of these changes will be the active setting.
When a project admin changes the default Code as Previous setting, users that have their Code as Previous settings changed will receive a toast notification and see a banner in their Code as Previous settings indicating that their settings have been changed by a project admin.
Code as previous project settings will be copied when using an existing project as a template for creating new projects. When using an existing project as a template, selecting to copy “Codes” settings will copy any project level Code as previous settings to the new project.
Learn more about code as previous and other project-level settings.
Allow empty prefix for PDF uploads without load file
You can now upload PDFs without a prefix using the PDF without load file uploader. By default, the EVER prefix will be given to PDFs without a valid Bates prefix. But you can now override this behavior by clicking the “Edit prefixes” button and selecting the “Allow Bates with empty prefix” option.
An “(empty prefix)” display value will be shown in the Bates prefix table. Note that a document name must start with a numerical digit in order for it to be assigned an empty prefix.
Learn more about Uploading Produced PDFs without a Load File.
Improved error reporting for Microsoft 365 Uploads
If a Microsoft 365 upload contains multiple custodians, your upload can still complete even if one or more custodians are unreachable. The upload card now displays a complete list of custodians whose data was attempted to be uploaded, and lists errors for any sources that could not be uploaded for each custodian.
The CSV file for this error displays the custodian ID, custodian name, and lists the failed Microsoft 365 source(s).
The CSV will automatically update with new errors while the upload is in process, so if a large upload takes many hours to complete, you can begin troubleshooting potential errors early. Of course, if you download the error report before the upload is complete, some errors may not yet be listed on the CSV.
Learn more about uploading documents from Microsoft 365 and troubleshooting upload errors here.
Improvements to legal holds directory table
The legal holds directory tab will now support pagination, sorting, and filtering. Users will also be able to filter and sort by each of the columns in the table, and they will be able to apply multiple filters at once. Users will have the option to issue a hold notice or export a CSV report of their filtered custodians.
To learn more about legal hold directories, please visit this article.
[Everlaw AI Beta] Review Assistant
The Review Assistant is comprised of tasks that can help you understand the content and relevance of documents faster and more comprehensively. Our initial tasks include:
- Document summary: Creates a narrative summary of the document
- Topic summary: Identifies distinct topics in the document and creates a summary by topic
- People and organizations extraction: Extracts references to entities in the text, and creates a brief description of the entity based on the text
- Dates and number extraction: Extracts references to numbers or dates in the text, and creates a brief description of the extracted number or date
- Sentiment analysis: Rates the document for positive, negative, or harmful or abusive sentiment, and provides a rationale for its rating based on the document
- Custom question and answer: Performs a task or answers a question based on user input
Review Assistant tasks can be accessed from the context panel. The results of the Assistant are not only helpful for understanding the content of a document, but can also be used as navigation aides to hone in on the most relevant parts of a document. For example, with the topic summary, you can highlight the document text relevant to a given topic.
For more information about Review Assistant, please see this support article.
[Everlaw AI Beta] Writing Assistant
The Writing Assistant can help you jumpstart writing tasks that rely heavily on document analysis. Not only does the Writing Assistant produce output based on parameters you set, it also clearly and transparently cites to relevant evidence. This makes it easy to verify the source and accuracy of information produced by the Assistant.
The Writing Assistant can be accessed from the AI icon in the draft toolbar. From there, you can choose whether you want to generate a statement of facts or a more custom piece of writing, and select which documents the Assistant should use when creating the draft.
For more information about the Writing Assistant, please see this support article.