PromptABCD
FeaturesLearnHow it worksUse casesFAQGuideBlogContext Blocks
Sign inGet started free
Sign inSign up
PromptABCD

A calm home for your best AI prompts. Save them once, find them in seconds, reuse them forever.

Product

  • Features
  • Free Courses
  • How it works
  • Use cases
  • Blog
  • Context Blocks
  • Export Anywhere
  • FAQ

Resources

  • User guide
  • Learn prompting
  • Sign in
  • Get started free

© 2026 PromptABCD. All rights reserved.

Privacy PolicyTerms and Conditions
Home/Blog/Gemini Prompts/Gemini Prompts for Legal Work
Gemini Prompts

Gemini Prompts for Legal Work

It's 9pm and you need an NDA draft for tomorrow's meeting — but a generic prompt can miss the one clause your client actually needs. These gemini legal prompts fix that with checklist-driven drafting.

July 22, 2026·8 min read
ShareShare
⚡Featured Prompt— copy and use right now
Write a standard NDA for a business meeting.

Picture this: you're a solo practice attorney, it's 9pm, and you need a first draft of a standard NDA for a client meeting tomorrow morning. You paste in a request, get something back that looks professional, and almost send it before noticing it's missing a mutual confidentiality clause your client specifically asked for. This is exactly where most gemini legal prompts go wrong — not because the output is unusable, but because the prompt didn't specify what actually mattered for this specific document.

Before going further: everything here is about drafting and organizing legal document language faster, always subject to review by a licensed attorney before use. Gemini doesn't practice law, doesn't know your jurisdiction's current case law, and shouldn't be the thing making a legal judgment call. This is about the drafting layer underneath that judgment.

Before: The Weak Prompt

Write a standard NDA for a business meeting.

What this does (poorly): produces a generic template that could apply to nearly any two parties, with generic clause language that doesn't reflect this specific client's actual concerns — mutuality, duration, carve-outs for information already public — because none of that was specified. The output looks like a real NDA. It just isn't the right one for this situation.

Why It Fails

Legal documents fail silently when they're generically correct but specifically wrong. An NDA that's one-directional when the client needed mutual protection isn't obviously broken — it reads fine, it's properly formatted, and a client without legal training might not catch the gap until it matters. The failure mode with vague legal prompts isn't garbled output; it's plausible-looking output that's missing the one clause that actually mattered for this deal.

⚠️ Common mistake: Treating "write me a [document type]" as a complete prompt because the output looks professionally formatted. Formatting quality and substantive correctness are completely different things, and AI-generated legal drafts are consistently strong on the former while needing real scrutiny on the latter.

After: The Improved Prompt

Draft a mutual NDA for [PARTY A] and [PARTY B], for the purpose of [SPECIFIC PURPOSE — e.g. "evaluating a potential joint marketing partnership"].

Required elements:
- Mutual obligations (both parties are disclosing and receiving confidential info)
- Term: [X years] with confidentiality surviving termination for [X years]
- Standard carve-outs: information already public, independently developed, or required to be disclosed by law
- Governing law: [STATE/JURISDICTION]

This is a first draft for attorney review, not a final document. Flag any section where jurisdiction-specific language would typically need customization.

What this does: specifies the actual deal-specific requirements (mutuality, term, jurisdiction) instead of relying on Gemini to guess what a "standard" NDA should include for this particular relationship, and explicitly frames the output as an attorney-review draft rather than something ready to send as-is.

⚡ Pro tip: The "flag any section needing jurisdiction-specific customization" instruction is worth including on every legal drafting prompt. Contract enforceability details vary meaningfully by state and country, and having Gemini proactively flag likely customization points — even without giving legal advice on what the customization should be — makes the attorney review pass faster and more targeted.

Breaking Down Each Element

The mutuality specification matters because it changes the entire structure of the document, not just a clause within it — a one-way NDA and a mutual NDA have different obligation structures throughout, and generating the wrong one from an ambiguous prompt means starting over rather than editing.

The explicit carve-outs list matters because these are the clauses clients most often assume are included by default and are most upset to discover missing — "why would I be liable for disclosing something that was already public" is exactly the kind of gap a generic prompt leaves open.

A paralegal at a small business law firm builds every contract drafting prompt around a firm-maintained checklist of required elements for each document type, rather than trusting a generic request to surface everything relevant. This turned contract drafting from "hope Gemini included what we need" into "verify Gemini included everything on our checklist," which is a meaningfully faster and more reliable review process.

⚡ Pro tip: Build a standing checklist of required elements for your firm's most common document types (NDAs, service agreements, employment offer letters) and paste the relevant checklist into every drafting prompt for that document type. This turns an open-ended "did it include everything" review into a fast pass-fail check against a known list.

Variations for Different Contexts

A real estate attorney uses a similar structured approach for lease agreement drafts, specifying the exact required elements for their state's landlord-tenant law (notice periods, security deposit handling, specific disclosures) rather than trusting a generic "residential lease" prompt to include jurisdiction-specific requirements automatically.

An in-house counsel at a mid-size software company uses the structured approach for vendor contract review rather than drafting — feeding in the vendor's proposed contract alongside the company's standard required-terms checklist and asking Gemini to flag every place the vendor's language deviates from the checklist, which turns contract review into a fast comparison task rather than a full manual read-through for a first pass.

⚠️ Common mistake: Skipping the "first draft for attorney review" framing because it feels redundant when you already know a human will review it. Making that framing explicit in the prompt itself changes the output — Gemini tends to flag more potential customization points and hedge appropriately on jurisdiction-specific language when it's told upfront that a legal professional will vet the result, compared to when it's asked to produce something that sounds finished.

Save and Reuse This

The reusable structure here is the checklist-driven prompt: specify the deal-specific requirements explicitly, attach your firm's standing checklist for that document type, and frame the output as an attorney-review draft rather than a finished product. New client, new deal terms, same reliable structure underneath.

Once your firm has built checklist-driven prompts for your most common document types, keep them organized and versioned rather than rebuilt from memory by whichever associate is drafting at 9pm before a morning meeting. PromptABCD gives legal teams a shared, versioned prompt library for exactly this, so the NDA checklist that caught a missing mutuality clause once doesn't get forgotten the next time someone's in a hurry.

More Scenarios From Legal Practice

A family law attorney uses a checklist-driven prompt for drafting initial separation agreement outlines, being explicit about which state's specific requirements apply to asset division and custody language, since these vary significantly and a generic template risks omitting a jurisdiction-specific requirement that a client would only discover was missing during a contested proceeding. She's careful to frame every output as a discussion starting point for the client consultation, never as something a client should sign without her full review and explanation.

A corporate paralegal at a venture-backed startup uses the structured approach for drafting board consent forms and equity grant documentation, maintaining a checklist tied to the company's specific cap table structure and prior board-approved templates, so each new document stays consistent with the company's established governance documents rather than introducing subtle variations that could create inconsistency across the company's legal history.

An immigration law firm uses checklist-driven prompts for drafting supporting documentation letters, with a checklist built around the specific visa category's documentary requirements. The attorney overseeing the practice group specifically requires the "flag jurisdiction or category-specific customization" instruction on every prompt, since immigration categories have specific, frequently updated requirements that a generic drafting request would have no way of knowing to include without being told explicitly.

Why the Checklist Habit Matters More in Legal Work Than Most Fields

Legal documents carry a specific kind of risk that differs from most other business writing: a missing clause or an outdated jurisdiction-specific requirement often isn't discovered until the document is tested — in a dispute, an audit, or a regulatory review — long after anyone would think to double-check it. This is different from, say, a marketing email with a weak subject line, where the failure is immediate and visible. The delayed, silent nature of legal drafting gaps is exactly why building a checklist into the prompt itself matters more here than in almost any other professional use case: the review has to happen before the document is used, because there may be no natural moment afterward that surfaces the gap until it's genuinely too late.

⚠️ Common mistake: Assuming that because a document type feels routine — an NDA, a standard service agreement — the drafting requirements are equally routine and don't need a jurisdiction-specific checklist. Even "boilerplate" documents have jurisdiction-specific enforceability quirks that a generic prompt has no way of accounting for on its own.

None of this makes the drafting process slower in any meaningful sense — a checklist pasted into a prompt takes seconds. What it does is turn the attorney review step from an open-ended "did we miss anything" search into a focused check against a known list, which is both faster and considerably harder to get wrong.

gemini legal promptslegal draftingcontract reviewnda templateslegal techai for lawyers

Continue Reading

Top 10 Gemini Tips Every User Should Know
Gemini Prompts

Top 10 Gemini Tips Every User Should Know

A forty-five minute manual reformatting job could have been a ten-second follow-up request. These gemini tips for users cover the habits that separate casual use from genuinely efficient work.

July 23, 2026·8 min read
Gemini for Startup Founders: Best Prompts
Gemini Prompts

Gemini for Startup Founders: Best Prompts

Most guides on gemini prompts for startup founders focus on pitch decks — but the real payoff is in prioritization. Here's how to use Gemini to catch the comfortable, low-risk work that quietly derails early-stage progress.

July 23, 2026·8 min read
Gemini for Academic Writing
Gemini Prompts

Gemini for Academic Writing

Can you use Gemini for academic writing without it counting as dishonesty? This case study shows exactly how one graduate student kept AI assistance on the right side of that line for her thesis.

July 23, 2026·8 min read

Save the prompts from this post

PromptABCD is a free prompt manager. Paste, organize, and reuse your best AI prompts — no more hunting through chat history.

Start free →
← PreviousGemini Prompts for Finance TeamsNext →Gemini Prompts for Real Estate Agents
Share this post:
ShareShare