Use-Case Inspiration for Using Everlaw's MCP Server Through a Connected AI Agent

Everlaw's MCP server allows you to securely leverage your Everlaw evidence from within a connected AI agent. When you bring your Everlaw evidence into the AI tool you're using for drafting and analysis, you set yourself up for deeper insights, faster.

This article describes use cases and scenarios where it's helpful to combine Everlaw with an AI agent. The goal is to provide guidance for those getting started with our MCP server, and to inspire the next step for those who are already using it.

Requirements

You must form the connection between Everlaw and your preferred AI tool before you can start. This typically involves some setup from an admin. Follow the instructions in the setup steps for the partner you'll be using. 

For more details on connecting, see Set Up Everlaw MCP Connectors.

Overall tips

These tips are not use-case specific, but will generally help you get and validate the answers you're looking for.

• Name the project: Start by telling the agent which Everlaw project to use, either by name or by pasting the project’s URL from your browser
• Start by asking about the project of interest: When you start by asking the agent to list the  binders, codes, productions, etc. on a project, both you and  your AI agent gain context about the matter that can be used to structure your searches
• Give the agent context: A short description of what you know about the case improves results: tell it the parties, key issues, important people, and relevant dates
• Ask the agent to show its work: Ask the agent to explain the steps it took, the tools it used,  and to cite the documents behind each conclusion. This makes it easier to check whether an answer is grounded in your evidence. If the first answer isn’t quite right, tell the agent what to change. If it can’t do something, ask why
• Verify before you rely on it: AI agents can make mistakes. Open cited documents to confirm results before using them in work product
• Ask the agent: When in doubt, ask the agent for advice on how to get the types of answers or documents you are looking for

Quick wins

The quick wins in this section describe common scenarios where a simple question can get you results that might be tedious or impossible without an AI agent.

Search in plain English

The quickest win is to perform a standard search from within the AI tool. If you're already doing work within an AI tool, it's a straightforward next step to explain the documents you're looking for and let the AI agent run the Everlaw search for you. This saves time as compared to separately logging into Everlaw, running the search on the search page, and then somehow copying the results into the agent.

Example question: “In [Project], find emails from Alex Carter to anyone at BlockVault sent between March and June 2021 that discuss pricing. Show me the Everlaw search you ran and provide a link to it.”

Find a specific document fast

Sometimes you're looking for a specific document, but you have trouble getting the exact details into the Everlaw search page. With an AI agent, you can provide the details you know, and it will try to find the document(s) for you.

• Bates number or specific metadata retrieval: When you know a specific Bates number, ask an AI agent to retrieve it for you. This is likely a bit faster than using the Everlaw Bates/Control # search, and can add the value of giving you an instant summary or overview of the document. 

  • Example question: "In [Project], retrieve the document with Bates number ABC0000067. Summarize its contents and identify the author, recipients, and date. Include a link to the document in Everlaw." 

    Tip

    This can work for any specific metadata parameter(s) e.g. "Locate the email from Alex Carter to Natalie Rhodes dated May 11, 2021, and provide its Bates number and a link to it in Everlaw."

• Identify a mystery document: When you have a vague understanding or recollection of a document, but don't have exact keywords or metadata parameters, you can describe it to the AI agent for it to identify candidate documents. 

  • Example question: "I’m trying to locate a communication suggesting possible financial impropriety. I believe a senior executive sent it shortly before the company’s collapse in 2022. Identify candidate documents and briefly explain why each may be the one."

    Tip

    This works best when searching projects with Deep Dive enabled.

• Turn a screenshot into a search: AI agents can "read" the text in an image. If you have a screenshot or image with relevant text for a search, paste the image into an AI agent. Then, ask the agent to use the text in the image to search Everlaw for the relevant document(s).
  • Example question: Using the text in the attached image, locate the matching document in [Project] and provide its Bates number and a link to it in Everlaw.

Get oriented in a matter

When you're new to a matter that has been going on for a while, you can use an AI agent to help you get oriented, highlight active areas of review, and point you toward what you need to know.

• Get the high level details: Ask questions to understand the key players and concepts
  • Example questions: 
    • Rank the custodians in [Project] by document volume and identify each custodian’s most active date range.
    • What date range in [Project] has the most chat activity?

• Understand case activity: Get up to speed on what your team has been up to, and what it means for the case
  • Example questions:
    • List the binders in [Project], with the document count for each and a one-line description of its apparent purpose.
    • Summarize the principal issues addressed in documents added to [Binder name] in the last 30 days
    • List the productions in [Project], including production date, Bates range, and document count.
    • Identify the 100 documents most recently viewed in [Project] and summarize the principal issues they address.

Understand a production or document set

Use an AI agent to help you get a handle data dumps.

• Test and refine search terms: When you're dealing with large tranches of data and need to figure out the key search terms, ask the agent to run proposed terms against your project and suggest improvements
  • Example question: “Here are our proposed search terms for [issue]. Run each term in [Project], report the number of hits, and suggest revisions for any term that appears overbroad or returns few results.”
• Gap analysis: Find out if there are notable gaps in date ranges or other metadata
  • Example question: Analyze the metadata for the [Production/Binder name] and identify any gaps or anomalies in date ranges, custodians, or file types that may indicate an incomplete production.
• Custom visualizations and dashboards: A flexible way to get custom visualizations across multiple criteria, such as keyword and metadata
  • Example question: Create a chart showing the number of documents in [Project] containing the term “energy,” grouped by year.
• Production review: An AI agent can summarize an incoming production for you. The summary can help you prepare to tackle it.
  • Example question: Summarize the RFP 124 upload for me. I want to know the distribution of custodians, file types, and date ranges.  

Next level 

If you're already running basic Everlaw searches from within an AI tool, you can use these suggestions to level up and get even more out of the agents.

Learn technical details

Use your AI agent to help you learn about technical or scientific concepts, in the context of your case. 

Example question: “Locate documents in [Project] that reference [scientific concept] Based on how the term is used in those documents, research it using reputable public sources and explain what it means and why it may be significant to the claims or defenses in this matter. Distinguish clearly between information drawn from the documents and information drawn from outside sources, and cite each."

Witness and deposition prep

• Public vs. private statements: If Claude has web access, it can compare what a witness has said publicly with what's in the documents
  • Example question: Identify Alex Carter’s statements in [Project] concerning [topic]. Then identify his public statements on the same topic and note any inconsistencies, citing the source of each statement.
• Full depo prep document: Let an AI agent draft your depo prep
  • Example question: I am preparing to depose Natalie Rhodes. Using documents in [Project], prepare a deposition outline that covers her role and key relationships, a chronology of events she was involved in, potential exhibits with Bates numbers and links, and proposed lines of questioning.

Let AI cite evidence for you

Precise citations are a requirement for legal writing, but they can be tedious to fill in. An AI agent connected to Everlaw can use your descriptions to identify the specific matching documents and link them in. This can be useful in:

• Memos
• Meet-and-confer letters
• Motion for Summary Judgment

Example question: For each factual assertion in this draft memo, identify supporting documents in the Key Evidence binder and add a citation by Bates number with a link to the document in Everlaw. Flag any assertion you cannot support.