Eve Agents

Last updated: March 10, 2026

Overview

Eve Agents are configurable digital teammates that perform case work — drafting documents, generating analysis, and executing tasks — either on demand or automatically. Agents replace and unify Eve's previous Playbooks and Blueprints into a single concept with expanded capabilities, including the ability to run automatically on a schedule or when triggered by case activity.

How It Works

Every agent has two dimensions that define what it does and how it runs:

Dimension 1: What Kind of Agent Is It?

Agents are defined by their output type. There are three kinds:

  • Chat Agents produce conversational responses — analysis, summaries, alerts, and recommendations. These replaced Playbooks.

  • Drafting Agents produce formatted documents saved to the matter or exported. These replaced Blueprints, with the improvement that instructions can now be saved and reused automatically.

  • Discovery Response Agents produce structured responses to written discovery — interrogatories, requests for production, and admissions. This is a specialized type of Drafting Agent.

Dimension 2: How Is It Triggered?

Any agent type can be run in two ways:

  • Manually — a user initiates the agent on demand, exactly like running a Playbook or Blueprint in the past. Manual execution is always available.

  • Automatically (via a Trigger) — the agent runs on a schedule (monthly) or in response to matter activity (such as a new document upload) with no user action required. Agents configured this way are called Auto Agents.

Any agent can be run manually or automatically — or both. Setting up a trigger does not remove the option to run the agent manually.

Migration from Playbooks and Blueprints

Existing Playbooks and Blueprints are automatically relabeled as Agents. No user action is required, and no functionality is lost. Blueprints that were set up with examples will still include those same examples after migration. The only change is terminology — Playbooks become Chat Agents, and Blueprints become Drafting Agents or Discovery Response Agents.

How Agents Relate to the Auditor

Agents and the Auditor are complementary capabilities, not competing features:

  • The Auditor surfaces insights — it flags risks, deadlines, and opportunities across cases. It identifies what matters need attention.

  • Agents execute work — they create drafts, produce analysis, generate discovery responses, and take action. They perform the case work.

The relationship is: Auditor identifies → Agent acts. For example, the Auditor may surface that a case is approaching policy limits, and a Demand Drafting Agent can be automatically triggered to begin drafting the demand letter.

Key Terms

  • Agent: A configurable instruction that tells Eve what to do (output type + instructions) and when to do it (manual, scheduled, or triggered).

  • Auto Agent: Any agent configured to run on a schedule or trigger without manual initiation. Also called an Automatic Agent.

  • Chat Agent: An agent that produces conversational output — analysis, summaries, alerts, or recommendations. Equivalent to Playbooks.

  • Drafting Agent: An agent that produces a formatted document or discovery response. Equivalent to Blueprints, with saved instructions as a new capability.

  • Discovery Response Agent: A specialized Drafting Agent that produces structured responses to written discovery.

  • Manual Agent: Any agent run on demand by a user. This is the default execution mode.

  • Central Library: A firm-wide document repository in Eve. Agents can reference Central Library files as context for drafting or analysis.

  • Trigger: A configuration that causes an Auto Agent to run automatically based on a schedule or matter activity.

  • Medical Overview: A built-in Eve capability that summarizes medical records. It is not currently configurable as an Auto Agent but can be run manually at any time.