Legal prompts
Prompts for first-draft clauses, plain-language contract summaries, and spotting risky terms — drafting aids, not legal advice.
Important upfront: the prompts in this category produce drafting aids, not legal advice. They help you generate a clean first draft of a mutual NDA, core clauses for a master services agreement, a statement of work with acceptance criteria, a consulting agreement, a demand letter, or a starter privacy policy — and they help you understand documents others send you, with prompts that summarize a contract in plain language, spot one-sided terms, or explain what each redline actually changes in practice. They are a starting point that a qualified attorney should review before anything is signed or sent.
Where AI genuinely helps is speed and comprehension. It's good at producing a structured first draft with consistent defined terms, at translating dense contract language into something a non-lawyer can act on, and at flagging risky clauses so you know what to ask about. That turns a blank page or an intimidating PDF into something workable.
Where it's dangerous is in the details: a model can invent a citation, miss a jurisdiction-specific requirement, or produce a confident clause that doesn't say what you think it says. So lean on the well-designed prompts here — the ones that flag assumptions, refuse to fabricate citations, and end with a counsel-review reminder — and never paste confidential or privileged information into a public AI tool, where your inputs may not stay private.
What makes a good legal prompt
A strong legal prompt is honest about its limits and disciplined about facts. It instructs the model to flag every assumption it's making, to use placeholders rather than invent party names, dates, or dollar figures, and to never fabricate case citations or statutes. It keeps defined terms consistent throughout the document, because an NDA that defines 'Confidential Information' three different ways is worse than no draft at all.
The best prompts also build in the guardrail directly: they ask for a plain-language explanation of what each clause does and why it matters, and they close with an explicit reminder that the output needs attorney review before use. That keeps the tool firmly in the 'draft and understand' lane and out of the 'rely on this' lane.
Get sharper results
- 01Never paste confidential, privileged, or personally sensitive information into a public AI tool. Use generic placeholders for names, terms, and figures, and fill the real details in privately afterward.
- 02Tell the model explicitly not to invent citations, statutes, or case law. If a draft cites authority, verify every reference independently — fabricated citations are a well-documented failure mode.
- 03Use the plain-language and redline-explanation prompts to understand a document before you negotiate it, but treat the risk flags as a list of questions for your attorney, not as conclusions.
- 04Always route a generated agreement through a qualified lawyer in the relevant jurisdiction before signing. Jurisdiction-specific rules — employment, privacy, leases — are exactly where a generic draft goes wrong.
Common questions
Can I just use an AI-drafted contract as-is?
No. These prompts produce a first draft to save you time, not a finished, enforceable agreement. Laws vary by jurisdiction and a model can miss requirements or word a clause ambiguously. Have a qualified attorney review anything before it's signed.
Is it safe to paste a contract I received into an AI tool?
Be careful. Confidential or privileged documents shouldn't go into public tools where inputs may be retained. If you need a plain-language summary, redact identifying and sensitive details first, or use a tool with proper confidentiality controls.
Why do these prompts insist on flagging assumptions and adding review notes?
Because it keeps the tool in its lane. Flagged assumptions show you exactly where the draft is guessing, and the review reminder prevents anyone from mistaking a convenient draft for vetted legal advice. That discipline is what makes AI drafting useful rather than risky.
Draft a Mutual NDA First Draft With Standard Clauses
Produces a clean first-draft mutual NDA with standard clauses, flagged assumptions, and an attorney-review reminder.
Draft Core Clauses for a Master Services Agreement
Generates first-draft MSA core clauses with consistent defined terms, flagged assumptions, and a counsel-review note.
Draft a Statement of Work With Scope and Deliverables
Drafts a structured SOW with deliverables, milestones, and acceptance criteria, plus flagged assumptions and a review reminder.
Draft Independent Consulting Agreement Clauses
Creates first-draft consulting agreement clauses covering contractor status, IP, and termination with flagged risks.
Summarize a Contract in Plain Language for Non-Lawyers
Turns dense contract text into a plain-language summary highlighting obligations, dates, and risks for non-lawyers.
Spot Risky and One-Sided Terms in a Contract
Scans a contract for one-sided or high-risk terms and proposes fairer alternatives, prioritized by severity.
Draft a Demand or Notice Letter for a Dispute
Drafts a professional demand or notice letter stating facts, remedy, and deadline, with flagged assumptions.
Draft a Starter Privacy Policy and Terms of Service
Produces a starter privacy policy and terms of service draft with placeholders and flags for regional legal tailoring.
Draft a Cease-and-Desist Letter First Draft
Drafts a firm first-draft cease-and-desist letter without fabricated citations, with flagged facts to verify.
Review Commercial or Residential Lease Clauses
Explains and flags key lease clauses from one party's perspective, with negotiation questions and a review note.
Draft Key Clauses for an Employment Offer Letter
Drafts core employment offer clauses including comp, status, and IP, with flags for jurisdiction-specific rules.
Explain Contract Redlines and Their Practical Impact
Explains each contract redline in plain language, classifies changes, and suggests counters with a counsel-review note.