The Lawyer's AI Toolkit: 12 Legal Prompts That Actually Save You Time (Contracts, Due Diligence & Court Docs)

The Lawyer's AI Toolkit: 12 Legal Prompts That Actually Save You Time (Contracts, Due Diligence & Court Docs)

Legal work is often repetitive: reviewing the same clauses, checking the same registries, drafting similar motions. While AI won't replace your judgment, it can handle the heavy lifting—if you know how to ask. The problem? Generic prompts like "write a contract" yield generic (and often wrong) results. The solution? Purpose-built prompts that mirror how legal professionals actually think.

In this guide, I've assembled 12 practical prompts across three skill levels—basic, advanced, and expert. Each prompt is designed to be copied, adapted, and used immediately. I've also included real examples and tips on how to verify AI output, because in law, accuracy is non-negotiable.

Why Prompts Matter in Legal Work

AI models are statistical text generators. They don't „know“ the law; they predict what's likely based on training data. That's why a vague prompt gives you a vague answer, while a structured prompt gives you a structured, useful answer. For legal tasks, structure is everything.

A good legal prompt should:
- Specify the role (e.g., „You are a corporate lawyer“)
- Define the task (e.g., „review this contract for risks“)
- Provide context (e.g., jurisdiction, industry, parties)
- Request a specific format (e.g., table, bullet points, clause-by-clause)
- Include constraints (e.g., „only flag issues that are legally enforceable“)

Now, let's dive into the prompts.

Basic Prompts: Getting Started

These prompts are simple but effective. They're perfect for routine tasks where you need quick, reliable output.

1. Drafting a Non-Disclosure Agreement (NDA)

Task: Generate a one-way NDA tailored to a specific jurisdiction and purpose.

Prompt:

You are a corporate lawyer with 10 years of experience drafting commercial agreements. Draft a one-way non-disclosure agreement (NDA) for a technology company (Disclosing Party) sharing proprietary software code with a potential business partner (Receiving Party). The NDA should:
- Be governed by the laws of the State of Delaware, USA
- Include a 3-year confidentiality term
- Define "Confidential Information" to include source code, algorithms, and trade secrets
- Include mutual non-solicitation of employees clause
- Specify that the Receiving Party may not reverse engineer the software
- Include a liquidated damages clause for breach (to be filled later)
- Be written in clear, plain English, avoiding overly complex legalese

Return the full text of the NDA, with placeholder brackets for names, dates, and specific amounts.

Example result: The AI produces a complete NDA with standard clauses (definition, exclusions, obligations, term, remedies). It even includes a section for „Liquidated Damages“ with a blank for the amount. You'd still need to review it for compliance with local law, but it's a solid first draft.

2. Summarizing a Legal Document

Task: Extract key terms from a dense contract.

Prompt:

You are a legal assistant. I'm giving you the text of a software licensing agreement. Summarize it in a structured format:
- Parties involved
- Granted rights (scope, exclusivity, sublicensing)
- Payment terms (amounts, schedule, late fees)
- Term and termination conditions
- Warranties and disclaimers
- Limitation of liability (cap and exclusions)
- Indemnification obligations
- Governing law and dispute resolution

For each section, quote the relevant clause number from the original text. If any section is missing, state "Not addressed." Be objective and do not add your interpretation.

Example result: The AI outputs a markdown table with sections and clause references, e.g., Section 3.2 for payment. This is great for a quick review or for preparing a contract summary for your client.

3. Checking a Contract for Common Red Flags

Task: Identify risky clauses in a contract.

Prompt:

You are a contract risk analyst. Review the following contract text and flag any clauses that are commonly considered high-risk in commercial agreements:
- Automatic renewal clauses
- Indemnification without limits
- Limitation of liability that excludes all consequential damages
- Non-compete provisions that are overly broad
- Assignment clauses that allow the other party to assign without consent
- Choice of law/forum that is inconvenient for our client
- Liquidated damages that may be considered penalties

For each flag, explain why it's risky and suggest a negotiation alternative. Present your findings in a table with columns: Clause Reference, Risk Level (Low/Medium/High), Risk Description, Suggested Change.

Example result: The AI gives you a table like this:

This is a great starting point for negotiation.

Advanced Prompts: Deeper Analysis

These prompts require more context and produce more nuanced output. They're for when you need a second opinion or a detailed analysis.

4. Due Diligence on a Company

Task: Analyze a company's legal health based on provided documents.

Prompt:

You are a due diligence specialist for an M&A transaction. I will provide you with the following documents for a target company: [list them, e.g., incorporation certificate, shareholder agreement, employment contracts, IP registrations, litigation history]. Analyze the legal risks and provide a due diligence report covering:
1. Corporate structure and governance (shareholder rights, board composition, voting agreements)
2. Intellectual property (ownership, licenses, potential infringement)
3. Employment (key employee contracts, non-competes, benefits)
4. Litigation and disputes (ongoing or potential)
5. Regulatory compliance (industry-specific licenses, permits)

For each area, list: (a) findings, (b) risk level, (c) recommended action items. Be thorough and cite document sections where applicable.

Example result: The AI produces a structured report with risk ratings and action items. For instance, it might note that the company's IP is owned by the founder personally, not the company—a common issue. You'd then recommend an IP assignment agreement.

5. Drafting a Statement of Claim for Court

Task: Generate a legal pleading based on facts.

Prompt:

You are a litigation attorney. Draft a Statement of Claim (complaint) for a breach of contract case in the [Court Name, e.g., U.S. District Court for the Southern District of New York]. The plaintiff is [Name], a [brief description], and the defendant is [Name]. The facts are as follows: [Provide a factual narrative, including dates, promises, and alleged breach]. The claim is for breach of contract, seeking damages of [amount]. Include:
- Jurisdiction and venue allegations
- Statement of facts in chronological order
- Count I: Breach of Contract (elements: existence of contract, performance by plaintiff, breach by defendant, damages)
- Count II: [Optional second claim, e.g., unjust enrichment]
- Prayer for relief (specific damages, interest, costs, attorney's fees)

Use proper legal formatting (e.g., "COMES NOW the Plaintiff...") and cite relevant legal standards (e.g., "To state a claim for breach of contract under New York law, a plaintiff must allege...").

Example result: The AI produces a draft complaint with the standard caption, numbered paragraphs, and legal citations. It even includes a note that you should verify court rules for formatting (e.g., local rules for page limits). This is a huge time-saver.

6. Analyzing a Court Decision

Task: Summarize a court opinion and its implications.

Prompt:

You are a legal analyst. I'm providing you with the text of a court decision (or a summary). Analyze it and provide:
- Case name and citation (if available)
- Procedural history (how the case got to this court)
- Facts: key facts that led to the dispute
- Legal issues: the specific questions of law the court addressed
- Holding: the court's decision on each issue
- Reasoning: the court's rationale, including key precedents cited
- Concurrences/dissents: if any, their main points
- Practical implications: how this decision might affect future cases or businesses

Write in plain English, but use legal terms where appropriate. If any information is missing, state that.

Example result: The AI outputs a structured analysis with headings like „Facts“, „Issue“, „Holding“. This is perfect for a legal memo or a blog post.

Expert Prompts: Complex and Multi-Step Tasks

These prompts are for complex tasks that require the AI to act as a strategic advisor or to handle multiple documents.

7. Creating a Risk Matrix for a Multi-Jurisdiction Deal

Task: Identify legal risks across multiple jurisdictions.

Prompt:

You are a cross-border M&A lawyer. I'm advising a client on acquiring a company with operations in the US, Germany, and China. Based on the following due diligence reports for each jurisdiction, create a comprehensive risk matrix. For each risk, include:
- Risk description
- Jurisdiction(s) affected
- Likelihood (Low/Medium/High)
- Impact (Low/Medium/High)
- Suggested mitigation strategy

Prioritize risks by combining likelihood and impact into an overall score. Provide a summary table and then a detailed narrative for the top 5 risks.

Example result: The AI generates a matrix with rows like „IP transfer restrictions in China“ and „German works council co-determination“. It scores them and suggests mitigation like „obtain CFIUS clearance“ or „restructure deal to avoid works council veto“. This is a powerful strategic tool.

8. Generating a Negotiation Playbook

Task: Develop a negotiation strategy for a contract.

Prompt:

You are a seasoned deal negotiator. Based on the following contract draft and the other party's likely interests, create a negotiation playbook. For each clause in dispute (list them, e.g., indemnification, liability cap, non-compete), provide:
- Our client's ideal position
- Walk-away position (the least acceptable)
- Likely counter-position from the other side
- Suggested compromise (e.g., „cap liability at 100% of fees, but include IP infringement carve-out“)
- Key talking points and supporting arguments
- Potential concessions to trade

Also identify any „red lines“ that should not be crossed. Present in a table format for each clause.

Example result: The AI produces a playbook like:

This gives you a roadmap for the negotiation.

9. Drafting a Merger Agreement Outline

Task: Create a detailed outline for a complex transaction.

Prompt:

You are a corporate lawyer specializing in M&A. Draft a detailed outline for a merger agreement between [Buyer] and [Target], both [jurisdiction] companies. The transaction is a [stock purchase / asset purchase / merger]. Include the following sections:
- Preamble and recitals
- Definitions
- Purchase and sale (including consideration, adjustments)
- Representations and warranties (for both parties, with survival periods)
- Covenants (pre-closing, post-closing)
- Conditions to closing
- Indemnification (including escrow/holdback)
- Termination and break-up fees
- Miscellaneous (notices, governing law, assignment)

For each section, provide a brief description of what should be included, and list key negotiation points. Do not draft the full text; just the outline with bullet points.

Example result: The AI gives you a comprehensive outline that you can use as a checklist. It might include notes like „consider MAC clause“ or „specify whether indemnification is subject to a basket and cap“. This is a great starting point for drafting.

10. Performing a Compliance Check Against GDPR

Task: Assess a contract's compliance with GDPR.

Prompt:

You are a data protection officer. Review the following contract (as a data processing agreement) for compliance with the EU General Data Protection Regulation (GDPR). Specifically, check for:
- Clear identification of the parties as data controller and data processor
- Description of the processing activities (nature, purpose, duration, categories of data)
- Obligation for the processor to follow documented instructions
- Confidentiality obligations
- Security measures (Article 32)
- Sub-processing rules (prior authorization)
- Data subject rights assistance (Article 28(3)(e))
- Data breach notification assistance
- Deletion or return of data after end of services
- Audit rights

For each requirement, state whether the contract addresses it, and if not, suggest a clause. Rate overall compliance as high/medium/low.

Example result: The AI produces a compliance checklist with a rating. It might flag missing audit rights and suggest adding a clause permitting the controller to conduct audits with reasonable notice. This is critical for EU operations.

11. Predicting Litigation Outcomes (with Caveats)

Task: Assess the strength of a case based on facts.

Prompt:

You are a litigation risk analyst. Based on the following facts and legal precedents, assess the likelihood of success for the plaintiff in a breach of contract claim. Provide:
- Summary of the legal elements
- Factual strengths and weaknesses for each element
- Relevant precedents (cite actual case law, if possible)
- Estimate of probability of success (low/medium/high) with reasoning
- Potential defenses and counterarguments
- Recommendation on whether to settle or proceed

Stress that this is not legal advice and that a human attorney must review.

Example result: The AI gives a balanced analysis, noting that while the plaintiff has strong evidence of breach, the defendant might argue waiver or statute of limitations. It might cite cases like Jacob & Youngs v. Kent for substantial performance. This helps you prepare for litigation strategy.

12. Automating a Contract Review Workflow

Task: Create a reusable prompt template for reviewing contracts.

Prompt:

You are a legal ops consultant. Create a standardized prompt template for reviewing vendor contracts. The template should:
- Define the role (e.g., "You are a legal reviewer")
- Include placeholders for contract text, jurisdiction, and specific concerns
- Output a structured report with sections for: key terms, risks, and recommended changes
- Include a rating system (e.g., red/yellow/green)
- Be concise but comprehensive

Provide the template in a copy-paste format.

Example result: The AI gives you a template like:

You are a legal reviewer. You will analyze the following contract: [Insert text]. The governing law is [Insert].

Focus on: [e.g., data protection, indemnification, termination].

Output a report with:
- Key terms (table)
- Risk assessment (red/yellow/green) for each clause
- Recommended changes (bullet points)
- Overall recommendation (approve/revise/reject)

You can then use this template repeatedly, making your workflow faster.

Ensuring Accuracy: The Human-in-the-Loop

AI is a powerful tool, but it's not infallible. Legal AI outputs can contain hallucinations—confident but false statements. Always verify:
- Case law citations: Check on official databases like Westlaw or LexisNexis.
- Contract clauses: Review for legal enforceability. AI might generate a clause that's invalid in your jurisdiction.
- Regulatory references: Consult official sources like eCFR for US regulations or EUR-Lex for EU law.

According to a Stanford study (2024), AI models hallucinate up to 20% of the time in legal tasks. So, treat AI as a junior associate—it's fast, but you must supervise.

Conclusion

The 12 prompts above cover a wide range of legal tasks, from drafting to analysis. By using them, you can save hours on routine work and focus on high-value strategy. But remember: AI is a tool, not a replacement for professional judgment. Always review and validate AI output before using it in any legal matter.

If you found this helpful, check out Asibiont.com for more AI-powered productivity guides. And if you have a favorite legal prompt, share it in the comments—I'd love to see how you're using AI in your practice.

Disclaimer: This article is for informational purposes only and does not constitute legal advice.