Skip to main content
Libraries are curated collections of reusable inputs that help Epiq AI workflows work more efficiently and consistently. Libraries provide known information up front so workflows can make better decisions with less manual correction. Some benefits of using libraries are:
  • Improved accuracy. Known entities are consistently identified and classified.
  • Reduced manual correction. Fewer errors to fix downstream.
  • Faster processing. Pre-identified entities speed up normalization and labeling.
  • Consistency across matters. Maintain entity information across multiple workflows.
Epiq AI supports four types of libraries: person, organization, prompt, and custom.

​
Library types

​
Person

Person libraries contain known individuals that are relevant to your data, such as attorneys or other recurring participants. Person libraries help workflows:
  • Identify people consistently across documents.
  • Normalize name variations into a single entity.
  • Apply roles or classifications more accurately where supported.
Use person libraries when you already know who is likely to appear in the data. Person entries carry a name (required), plus optional email, name and email aliases, organization, job title, first/middle/last name, start and end dates, description, location fields, and tags.

​
Organization

Organization libraries contain known organizations or domains, such as law firms, client companies, vendors, or opposing parties. Organization libraries help workflows:
  • Recognize organizations across metadata and content.
  • Normalize organization name and domain variations.
  • Maintain consistent handling of organizations across workflows and matters.
Use organization libraries when documents involve repeat companies or firms. Organization entries carry a name (required), plus optional domain and domain aliases, name aliases, industry, location, country, state, city, description, and tags.

​
Prompt

Prompt libraries store reusable prompt content that can be referenced by workflows. Prompt libraries help you:
  • Standardize prompt language across workflows.
  • Reuse approved or validated prompt text.
  • Reduce duplication and manual prompt management.
Use prompt libraries when multiple workflows rely on the same or similar prompt instructions. Prompt entries carry the prompt text (required), plus an optional title and tags.

​
Custom

Custom libraries hold reference data that does not fit the person, organization, or prompt shapes — a list of matter codes, deal names, product identifiers, custodian groupings, or any other lookup table your matter depends on. Unlike the other three types, a custom library has no fixed columns. You define them when you create it, and the columns you define become the library’s import template, its table view, and what you can filter and search on. Use custom libraries when you have reference data worth reusing across projects that the built-in types cannot represent.

​
Library scope

Every library belongs to your account. Beyond that, scope is determined by what you associate it with, and Epiq AI labels each library accordingly on the Libraries page: A library associated with a client organization takes precedence in this labeling: it reads as a client library even when it is also applied to individual projects.

​
Client libraries

A client organization is a contact assigned to a project as its client. Associating a library with one or more client organizations makes it a client library — it becomes available across every project for that client, present and future, without being applied to each one individually. This is the right scope for reference data that belongs to the client rather than to a single matter: the client’s own attorneys and in-house counsel, its subsidiaries and domains, its regular outside firms, or standing prompt language the client has approved. Two behaviors follow from the association:
  • When you create a library from inside a project, and you do not choose client organizations yourself, Epiq AI associates it with that project’s client automatically.
  • A library that carries client organizations can only be applied to projects belonging to those clients. Attempting to apply it elsewhere is refused with “The library belongs to a different client.” This is what keeps one client’s reference data out of another client’s matters.
Leave the client organizations field empty to keep a library global — the field’s own helper text says “Global - available to all.”

​
Applying a library to projects

A library does nothing on its own; it has to be applied to the projects that should use it. You can apply libraries when you create them, or later from the library’s detail panel:
  • The Projects tab lists the projects the library is applied to, and lets you assign or unlink them.
  • The Client Orgs tab manages which client organizations the library is associated with.
  • The Details tab holds the name and description.
Unlinking a library from one project leaves every other project’s use of it untouched.

​
How workflows use libraries

Libraries are read by workflow components rather than being applied by hand:
  • Enrich from Libraries pulls known information into the entities a workflow found, so a person the library already describes arrives with that description.
  • When an entity a workflow identifies already exists in a library, its tags are copied directly rather than inferred again. This both speeds the run up and keeps tagging consistent with what has already been verified.
  • Save to Libraries writes newly identified entities back, so what one matter established is available to the next.
See Entity Labeling for how library-sourced tags feed privilege decisions.