When to Use Ask vs Interact vs Draft

Lexlegis.ai offers three core workflows, Ask, Interact, and Draft, each designed for a distinct stage of legal work.

Choosing the appropriate workflow ensures higher accuracy, efficient use of documents, and clarity across research, analysis, and drafting. This guide explains what each workflow is designed to do and when to use it.

1. Overview

Each workflow on Lexlegis.ai is defined by its purpose and starting point.

At a high level:

• Ask answers questions of law and interpretation
• Interact works directly on documents to extract insights
• Draft creates structured legal documents

2. Ask: Judicial Interpretation and Legal Research

Use Ask for end-to-end, source-grounded, hallucination-free responses to questions of judicial interpretation across the Lexlegis database, with clear source references.

Ask also supports research on user-uploaded files, folders, or knowledge bases, as well as controlled web-based legal research where specific websites can be included or excluded.

2.1 Typical Use Cases

• Judicial and statutory research
• Interpreting questions of law and questions of fact
• Researching across user-uploaded files and knowledge bases
• Conducting ringfenced legal research on the web

2.2 What Ask Is Designed For

• Producing structured legal reasoning grounded in authoritative sources
• Providing transparent citations to statutes, case law, and referenced materials
• Ensuring responses remain source-anchored and hallucination-free

Ask is most effective when the starting point is a legal or interpretive question.

3. Interact: Document Analysis and Actionable Insights

Use Interact to interact with individual or multiple documents and extract actionable insights from them.

Interact supports tabular and narrative analysis using preset or custom Question Sets, enables due diligence workflows in minutes, and allows the application of Lexlegis preset prompts for actions such as summaries, risk and review assessments, comparisons, and fact sheet generation.

3.1 Typical Use Cases

• Reviewing multiple contracts, agreements, or pleadings
• Extracting timelines, issues, or structured summaries
• Comparing documents or drawing inferences across files

3.2 What Interact Is Designed For

• Working with one or more documents simultaneously
• Applying prompts or Question Sets consistently across files
• Producing narrative or tabular outputs
• Clearly indicating which documents informed the analysis

Interact is most effective when the starting point is documents.

4. Draft: Creating Legal Documents

Use Draft to create legal documents quickly and flexibly, either from predefined templates or entirely from scratch using custom instructions.

Draft supports first-instance drafts and replies, allows the application of preset boilerplate clauses on an as-is basis, and enables the use of AI Drafting Rulebooks to define granular drafting standards and ensure consistency.

Once a draft is generated, it can be refined using in-line edits and AI-assisted actions such as tone adjustments or section-specific instructions. Draft also supports collaboration and maintains a version-controlled audit log.

4.1 Typical Use Cases

• Drafting contracts and agreements
• Replying to legal notices, motions, and orders
• Drafting documents using preset or custom templates

4.2 What Draft Is Designed For

• Generating structured legal documents aligned with defined drafting rules
• Applying clauses, boilerplates, and drafting standards consistently
• Producing editable drafts with version control and collaboration support

Draft is most effective when the starting point is a defined legal deliverable.

5. Choosing the Appropriate Workflow

The workflows on Lexlegis.ai are designed to work together and can be used sequentially within the same matter.

A simple way to distinguish them is by identifying the starting point:

• A legal or interpretive question leads to Ask
• A set of documents leads to Interact
• A required legal output leads to Draft

Moving between workflows as work progresses is expected and supported by the platform.