Tutorial 02: Getting Started with OpenAI for Legal Work
Learn how to choose the right OpenAI plan, structure effective legal prompts, and complete your first contract analysis and research workflow.
Learning Objectives
By the end of this tutorial, you will:
- Understand which OpenAI plan is right for your legal practice
- Know how to structure legal prompts effectively
- Complete your first contract analysis
- Set up a basic legal research workflow
Choosing Your OpenAI Plan
Available Options
| Plan | Price (Public List) | Best For | Key Features |
|---|---|---|---|
| Free | $0 | Testing/light use | Basic ChatGPT access |
| Plus | ~$20/mo | Solo practitioners | More usage, GPT-4, file uploads, Custom GPTs |
| Team | From $25/user/mo | Small firms | Team admin controls and shared workflows |
| Enterprise | Custom | Large firms | SSO, dedicated support, compliance |
Recommendation for Legal Professionals
If you need more than the free tier, ChatGPT Plus is the usual starting point. This gives you:
- More usage than Free
- GPT-4 access for document analysis
- File uploads for contract review
- Custom GPTs for matter-specific workflows
- Better day-to-day workflow depth for active legal practice
Upgrade to Team or Enterprise when you:
- Hit message limits regularly
- Need to share workflows with colleagues
- Require admin controls for compliance
Pricing references are vendor-published list prices for context only. Legalai.guide is free and independent. Always verify current pricing on OpenAI's official site.
Anatomy of an Effective Legal Prompt
Legal work requires precision. A well-structured prompt dramatically improves ChatGPT's output quality.
The CRISP Framework for Legal Prompts
C - Context (Who are you? What's the situation?) R - Role (What expertise should ChatGPT assume?) I - Instructions (What specific task?) S - Specifics (Jurisdiction, format, constraints) P - Parameters (Length, tone, citations needed?)
Example: Generic vs. CRISP Prompt
Generic Prompt (Avoid)
CRISP Prompt (Use This)
Your First Contract Analysis
Exercise 1: Indemnification Clause Review
Time: 10 minutes
Scenario: A vendor has sent you their standard SaaS agreement. Your client is the customer. Analyze this indemnification clause.
Step 1: Copy this prompt to ChatGPT:
Step 2: Review ChatGPT's response
Step 3: Ask these follow-up questions:
- "What's the most important missing provision from Vendor's indemnity?"
- "Draft a mutual indemnification provision that would be more balanced"
- "How would you prioritize these asks if we can only get 2 changes?"
What to Look For in ChatGPT's Response
Good signs:
- Identifies asymmetry (customer indemnifies more broadly)
- Notes missing data breach/security indemnity
- Recognizes HIPAA-specific risks
- Provides practical, negotiable language
Always verify:
- Specific legal citations
- State-specific requirements
- Recent case law references
Basic Legal Research Workflow
Understanding ChatGPT's Research Capabilities
What ChatGPT Can Do Well:
- Explain legal concepts and doctrines
- Identify relevant areas of law to research
- Synthesize information from documents you provide
- Generate research outlines and issue checklists
- Draft memoranda based on research you supply
What Requires Additional Tools:
- Current case law citations (use a dedicated case law database)
- Up-to-date statutory language (verify against official sources)
- Local rules and procedures (always verify)
Case Law Source Options (Alternatives)
Use one primary source and one backup source for citation verification.
Paid platforms:
- Westlaw / Lexis: Primary commercial research platforms
- vLex (incl. Fastcase Library): Commercial alternative with broad case law coverage
Free/public options:
- CourtListener (Free Law Project): Large free opinion database + RECAP archive
- Google Scholar (Case law mode): Fast, free starting point for case lookup
- Justia / Cornell LII: Helpful for quick checks and primary-law navigation
Exercise 2: Research Planning
Scenario: Your client wants to know if their non-compete agreement is enforceable.
Step 1: Use this research planning prompt:
Step 2: Review ChatGPT's research plan
Step 3: Follow up with:
- "What's the leading case on this conflict of laws issue?"
- "If we were representing the employer, what's our best argument?"
- "Draft an outline for a memo analyzing this issue"
Pro Tip: Using ChatGPT with Traditional Research
ChatGPT works best as a research accelerator, not a replacement:
- Start with ChatGPT: Get issue framework and search strategies
- Run searches in Westlaw/Lexis: Find actual citations
- Return to ChatGPT: Upload cases/statutes for synthesis
- Draft with ChatGPT: Create memo structure
- Review and verify: Always check citations yourself
Document Drafting Basics
The Three-Stage Drafting Process
Stage 1: Outline & Strategy
Stage 2: First Draft
Stage 3: Refinement
Exercise 3: Draft a Client Communication
Scenario: You need to explain a complex settlement offer to a client.
Setting Up Your Legal Workspace
Organizing with ChatGPT Memory and Custom GPTs
ChatGPT memory and Custom GPTs let you create persistent workspaces for different matters or practice areas.
How to Set Up a Custom GPT for Contract Review:
- Go to ChatGPT and create a new Custom GPT
- Add a clear name (e.g., "Acme Corp - MSA Review")
- Write custom instructions (your playbook)
- Configure file upload capability for relevant documents
Example Custom Instructions for Contract Review:
Recommended Structure for Law Firms
Create these standard Custom GPTs:
- [Client] - [Matter]: Matter-specific workspace
- Contract Playbook - [Type]: Your firm's standard positions
- Legal Research - [Practice Area]: Practice area knowledge base
- Templates & Precedents: Standard documents for reference
Quality Control Checklist
Before relying on any ChatGPT output:
Citation Verification
- Verify all case citations in Westlaw/Lexis
- Check that cited statutes are current
- Confirm procedural rules match your jurisdiction
- Validate any regulatory references
Accuracy Check
- Do the legal conclusions follow from the analysis?
- Are there counterarguments not addressed?
- Does the advice fit your specific facts?
- Have you checked for recent developments?
Client-Ready Review
- Remove any obvious "AI-generated" language
- Verify tone is appropriate for audience
- Check for any confidential information leakage
- Ensure advice is actionable and specific
Homework Before Tutorial 03
- Complete all three exercises in this tutorial
- Create your first Custom GPT for a current matter
- Draft custom instructions for one type of work you do regularly
- Try 5 different legal tasks with ChatGPT and note what works/doesn't
Quick Reference: Essential Legal Prompts
Contract Review
Legal Research
Document Drafting
Client Communication
Do This Now
- Choose your OpenAI plan and complete Exercise 1 (indemnification clause review)
- Run the research planning prompt (Exercise 2)
- Draft a client communication (Exercise 3)
Continue to Document Analysis -> | Previous: Overview
Related family pages
- Claude Getting Started - Same concepts with Claude
- Core Concepts - Platform-neutral legal workflow model
Sources
- OpenAI Pricing
- OpenAI API Docs
- Prompt Engineering Guide
- CourtListener
- Google Scholar
- Justia Law
- Cornell LII
- vLex