AI Prompts for NDA: 18 Templates Across ChatGPT, Claude And Gemini

An NDA is usually the first document two companies exchange, and it sets the tone for everything after it. Too aggressive and it signals distrust before the relationship starts. Too loose and it does not actually protect what either side cares about. Most teams either reuse a template that was written for a completely different kind of relationship, or spend an afternoon on a document that should take twenty minutes.

The 18 templates here, divided across ChatGPT, Claude and Gemini, treat an NDA as a document that should be fast to produce and precise about what it covers. The ChatGPT prompts cover mutual and one-way NDAs, plain-language summaries and quick clause rewrites. The Claude prompts handle messy internal requests, redlines against a standard, and the definition, term and multi-party sections that need to be exact. The Gemini prompts research jurisdiction norms, industry-standard terms and the regulatory context around confidential data. None of this replaces a lawyer reviewing anything before it is signed, but it gets the contract process moving with a strong first draft on the table fast.

Why AI Works Well For NDAs

An NDA is a short, structured document with a small number of clauses that matter enormously. AI handles that combination well when the brief specifies exactly what needs protecting and for how long, and it helps to know what actually goes into one before you start prompting.

Each Model Has A Different Edge

ChatGPT is the most flexible for standard mutual and one-way drafts. Claude handles messy requests and precise clause language best. Gemini is the strongest choice when jurisdiction norms or recent regulatory context matter.

Mutual Or One-Way Changes Everything

The single biggest drafting mistake is using a mutual NDA when only one side is disclosing anything, or vice versa. Decide this before writing a word, because it changes which obligations belong in the document at all.

Define Confidential Information Precisely

A vague definition of what counts as confidential is the clause most likely to cause a dispute later. Ask the model to be specific about what is covered and what is explicitly excluded, such as information already public or independently developed.

This Is A Drafting Starting Point, Not Legal Advice

Every template below produces a strong first draft, not a document ready to sign untouched. Have a qualified lawyer review any NDA before it goes to a counterparty, especially where jurisdiction, liability or enforceability questions are involved.

ChatGPT Prompts For NDAs

ChatGPT is the flexible workhorse for standard NDA drafting. It handles mutual and one-way agreements, quick summaries and fast clause rewrites, all ready for e-signature once the language is settled. These six ChatGPT prompts for NDA writing cover the situations that come up most often before a deal, a hire or a vendor relationship gets underway. Each ChatGPT prompt for NDA below is built around a specific scenario, so you can pick the right ChatGPT prompt for NDA for the job rather than starting from a blank page.

1. Full Mutual NDA From Scratch

Act as a commercial lawyer's drafting assistant producing a standard mutual NDA.
Context:
- Party A: [name and jurisdiction]
- Party B: [name and jurisdiction]
- Purpose of disclosure: [why the parties are sharing information]
- Term of confidentiality obligation: [e.g. 2-3 years]
- Governing law: [jurisdiction]
Write a complete mutual NDA.
Sections:
1. Definition of confidential information, with specific exclusions
(already public, independently developed, disclosed with consent)
2. Obligations of both parties
3. Permitted disclosures (advisors, as required by law)
4. Term and survival of obligations after termination
5. Return or destruction of information
6. No licence or ownership transfer implied
7. Governing law and jurisdiction
8. Signature blocks for both parties
Rules:
- Both parties have identical obligations, since this is mutual.
- Plain, professional legal language, not overly aggressive.
- Flag this as a drafting starting point requiring legal review.

Where it works best: ChatGPT produces a complete, correctly structured mutual NDA in one pass, with both parties’ obligations genuinely symmetrical. This is the ChatGPT prompt for NDA most teams reach for first.

Best for: Two companies about to exchange confidential information in both directions, such as a partnership or joint evaluation.

Turn a signed NDA into your first branded document in 60 seconds
Paste your website URL and Proposal.biz pulls your brand, services and case studies into a Smart Content Library, so the proposal or SOW that follows the NDA looks like yours automatically.
Try for free

2. One-Way (Unilateral) NDA

Act as a commercial lawyer's drafting assistant producing a one-way (unilateral) NDA.
Context:
- Disclosing party: [name]
- Receiving party: [name]
- What is being disclosed: [the nature of the information]
- Reason for disclosure: [e.g. evaluating a potential engagement]
- Term: [duration of the obligation]
Write a complete one-way NDA where only the receiving party has confidentiality obligations.
Sections:
1. Definition of confidential information, specific to what is actually being shared
2. Receiving party's obligations only
3. Standard exclusions (public information, independently developed)
4. Permitted disclosures
5. Term and post-termination survival
6. Remedies for breach, stated plainly
7. Governing law
8. Signature blocks
Rules: do not accidentally add mutual obligations. Only the receiving party is bound. Flag as a drafting starting point requiring legal review.

Where it works best: ChatGPT keeps the one-directional structure clean, which is the detail most templates get wrong by defaulting to mutual language out of habit.

Best for: Situations where only one side is actually disclosing anything, such as a vendor evaluating a client’s data before scoping work.

3. NDA Cover Email

You are writing the short email that sends an NDA to a counterparty for signature.
Context:
Counterparty: [name]
Relationship so far: [first contact, post-call, ongoing]
NDA type: [mutual or one-way]
Deadline, if any: [date]
Write a short, professional cover email.
Rules:
- Confirm what is attached and why.
- One sentence on what happens after signature (e.g. "we can then share the technical details").
- Keep it under 90 words.
- Warm but businesslike, not overly formal.
- End with a clear, low-friction next step.

Where it works best: ChatGPT writes a concise, appropriately toned email fast, which matters since an overly formal NDA cover note can feel more adversarial than the document itself.

Best for: The message that actually gets the NDA signed and returned quickly.

4. Plain-Language Summary Of An NDA

You are explaining a signed or draft NDA to someone without a legal background.
NDA text:
[paste the NDA, or the sections in question]
Task:
Write a plain-language summary of what this NDA actually means.
Output:
- What information is covered, in plain terms
- What each party is and is not allowed to do with it
- How long the obligation lasts
- What happens if it is breached, in plain terms
- Anything unusual or worth flagging compared to a standard NDA
Rules: no legal jargon. Write it the way you would explain it to a colleague who has never read a contract. This is not legal advice, only a plain-language summary.

Where it works best: ChatGPT translates dense clause language into something a non-lawyer stakeholder can actually act on before signing.

Best for: Getting internal sign-off from someone who needs to understand the terms but will not read the legal text closely.

5. Rewrite An Overly Aggressive NDA Clause

You are a contracts specialist who negotiates NDA terms that are fair to both sides.
Clause in question:
[paste the specific clause a counterparty sent, or our own draft clause]
What concerns us about it: [your concern, e.g. term too long, definition too broad, one-sided remedies]
Task:
Rewrite this clause to be fair and specific, while still genuinely protecting confidential information.
Rules:
- Keep the clause's legitimate purpose intact.
- Narrow anything overly broad (e.g. an unlimited term, an unreasonably wide definition of confidential information).
- Explain in one sentence why the rewrite is more reasonable.
- Flag that any counterparty-facing change should go through legal review before being sent back."

Where it works best: ChatGPT is quick at spotting why a clause reads as one-sided and proposing a fair middle ground without losing the clause’s actual purpose.

Best for: Negotiating a counterparty’s NDA draft rather than starting from your own template.

6. Vendor/Contractor NDA (Bonus)

Act as a procurement lead's drafting assistant producing an NDA specific to a vendor or contractor relationship.
Context:
- Our company: [name]
- Vendor/contractor: [name]
- What they will have access to: [systems, data, or materials]
- Duration of the engagement: [timeframe]
- Whether they will subcontract any of this work: [yes/no]
Write a one-way NDA tailored to a vendor or contractor context.
Sections: definition of confidential information (specific to what they will actually access), obligations, subcontractor flow-down clause if relevant, return of materials at engagement end, term, governing law, signature blocks.
Rules: be specific about system or data access rather than using generic language. Flag as a drafting starting point requiring legal review.

Where it works best: ChatGPT adapts the standard one-way structure cleanly to the specifics of a vendor relationship, including the subcontractor flow-down clause that generic templates often miss.

Best for: Onboarding a new vendor or contractor who will touch sensitive systems or data.

Claude Prompts For NDAs

Claude is the right model when the request behind the NDA is messy, or when a specific clause needs to be exact rather than approximate, which is exactly what legal teams need before anyone signs anything. These six Claude prompts for NDA writing handle the situations where precision matters most. Each Claude prompt for NDA below is built around a specific scenario, so you can pick the right Claude prompt for NDA for the job rather than starting from a blank page.

1. Full NDA From A Messy Internal Request

You are a contracts specialist who turns a rough internal request into a precise NDA.
[paste the internal request, email thread, or Slack message asking for an NDA, however unstructured]
What we typically require: [your company's standard NDA terms, if any]
Task:
Turn this request into a complete, correctly structured NDA.
Instructions:
- Identify whether this should be mutual or one-way from the request, and state your reasoning.
- Pull out what is actually being disclosed and define it specifically.
- Where the request is vague about term or scope, propose a specific default and label it as something to confirm.
- Produce the full NDA: definition, obligations, exclusions, term, remedies, governing law, signature blocks.
Flag as a drafting starting point requiring legal review before it is sent to any counterparty.

Where it works best: Claude reads unstructured requests faithfully and makes sound, clearly labelled default decisions where the request itself was vague. This is the Claude prompt for NDA most teams reach for first.

Best for: The common real-world case: someone asks for “an NDA for this deal” with almost no detail attached.

Know the moment your counterparty opens the NDA
Send the signed-ready NDA as a trackable link instead of a flat PDF attachment, and see exactly when it was opened, so you are never wondering whether to follow up.
Try for free

2. NDA Redline / Comparison Against A Standard Template

You are a contracts specialist comparing a counterparty's NDA draft against our standard template.
Our standard NDA: [paste our template, or its key terms]
Counterparty's draft: [paste their draft]
Task:
Compare the two and produce a redline summary.
Output:
- A table or list of every material difference: term length, definition scope, remedies, governing law, assignment rights
- For each difference, a one-line note on whether it favours us, them, or is neutral
- A recommended position for each difference worth pushing back on
Rules: be precise about what actually changed, not a general impression. Flag anything unusual compared to standard market terms.

Where it works best: Claude compares two dense legal documents accurately and holds the comparison structure cleanly across many clauses.

Best for: Reviewing a counterparty’s own NDA draft instead of sending your own template.

3. Confidentiality Definition Section

You are a contracts specialist writing the single most important clause in any NDA: what counts as confidential.
Context:
What is actually being shared: [describe it specifically]
What should explicitly NOT be covered: [e.g. information already public, independently developed, received from a third party without restriction]
Task:
Write a precise definition of confidential information.
Requirements:
- Cover the specific categories of information actually at stake, not a generic catch-all.
- Include the standard carve-outs explicitly, so the definition cannot be read to cover information that should not be restricted.
- Keep the language enforceable rather than sweeping.
Output the clause, then one paragraph explaining the reasoning behind each carve-out included.

Where it works best: Claude is precise about scope and carve-outs, which is exactly what keeps this specific clause from being either too narrow to protect anything or so broad it becomes unenforceable.

Best for: Any NDA where the information being protected is specific and technical enough that a generic definition would not hold up.

4. Term And Survival Clause

You are a contracts specialist writing the term and survival clause of an NDA.
Context:
Type of information: [e.g. trade secrets, standard business information, technical specifications]
Relationship duration: [how long the underlying relationship is expected to last]
Industry norm for this kind of information, if known: [your sense of what is typical]
Task:
Write a term and survival clause.
Requirements:
- State how long the NDA itself is in effect.
- State how long confidentiality obligations survive after termination or expiry, and explain briefly why that duration fits the type of information.
- Address what happens to obligations if the underlying relationship (e.g. a deal) does not proceed.
Rules: do not default to "indefinite" without a clear reason.
Match the duration to the actual sensitivity and shelf life of the information.

Where it works best: Claude reasons carefully about matching term length to the actual nature of the information, rather than defaulting to boilerplate durations.

Best for: Situations where an indefinite or overly long confidentiality term would be commercially unreasonable to agree to.

5. Multi-Party NDA (Three Or More Parties)

You are a contracts specialist drafting an NDA involving more than two parties.
Context:
Parties: [list all parties and their role, e.g. company, investor, advisor]
Purpose of disclosure: [why information is being shared among this group]
Direction of disclosure: [who is sharing with whom, if not fully mutual]
Task:
Write a multi-party NDA.
Requirements:
- Clearly state each party's role and obligations, since they may not be identical for everyone.
- Address how information shared by one party is protected from the others, not just from outside disclosure.
- Cover what happens if one party is added or removed later.
- Standard sections: definitions, obligations, term, governing law, signature blocks for all parties.
Flag as a drafting starting point requiring legal review, and note that multi-party structures particularly benefit from counsel review given the added complexity.

Where it works best: Claude handles the added complexity of asymmetric obligations across multiple parties without the structure collapsing into a confusing document.

Best for: Deals involving an investor, an advisor, or more than two companies at the table.

6. Tighten An Existing NDA Draft

You are a careful editor of legal drafting, working on an NDA.
Draft:
[paste the full NDA draft]
Task:
Tighten this draft without changing its legal substance.
Instructions:
- Simplify overly dense sentences without losing precision.
- Flag any clause that is genuinely ambiguous, rather than just wordy, since that is a substantive issue, not a style one.
- Note any clause that seems to be missing compared to a standard NDA (e.g. no survival clause, no governing law).
- Do not change obligations, terms, or scope, only clarity.
Output:
1. The tightened draft.
2. A list of anything flagged as ambiguous or missing, for legal review before this goes any further.

Where it works best: Claude is careful to separate genuine style tightening from substantive legal changes, and is explicit about flagging the difference rather than quietly altering meaning.

Best for: A draft that is legally sound but reads as dense or was assembled from mismatched clause templates.

Gemini Prompts For NDAs

Gemini’s live web grounding is the right tool when jurisdiction norms, industry-standard terms or recent regulatory context need to inform the NDA before it is sent, right down to confirming who actually has signatory authority on each side. These six Gemini prompts for NDA writing turn a generic template into one grounded in current context. Each Gemini prompt for NDA below is built around a specific scenario, so you can pick the right Gemini prompt for NDA for the job rather than starting from a blank page.

1. Research-Backed Jurisdiction-Appropriate NDA

You are a contracts specialist who checks jurisdiction norms before finalising an NDA.
Step 1: Research standard NDA practice and enforceability considerations in [jurisdiction] for [type of information, e.g. trade secrets, technical data].
Look for:
- Common term lengths considered reasonable in this jurisdiction
- Any statutory definition of trade secrets or confidential information that applies
- Common governing law and dispute resolution choices for this kind of agreement
Step 2: Write an NDA informed by what you find, for [party A] and [party B] under [jurisdiction] law.
Requirements: cite the source for any jurisdiction-specific claim.
If you cannot verify something, say so rather than guessing. Flag as a drafting starting point requiring local legal review.

This turns a generic AI prompt for NDA writing into one grounded in the actual jurisdiction's norms.

Where it works best: Gemini’s live web grounding surfaces jurisdiction-specific norms that a generic template would miss entirely. This is the Gemini prompt for NDA most teams reach for first.

Best for: Cross-border deals, or any NDA where the governing jurisdiction is not one you draft in every day.

Draft, send and e-sign the whole document chain in one window
Keep NDA, proposal and contract in one place: AI-assisted drafting, automatic branding, a trackable link to send, and built-in e-signature so nothing stalls waiting on a separate tool.
Try for free

2. Industry-Standard NDA Terms Benchmark

You are an analyst researching typical NDA terms in a specific industry, so ours does not look unusual to a counterparty.
Step 1: Research typical NDA terms for [industry, e.g. SaaS partnerships, M&A due diligence, healthcare vendor evaluation].
Find:
- Common term lengths
- Common scope of what is covered
- Any industry-specific carve-outs that are standard
Step 2: Write a short benchmark summary, then flag anything in our current draft that is unusually broad or narrow compared to the norm.
Our current draft or key terms: [paste them]
Output: a paragraph on the industry norm with sources, then a short list of anything in our draft worth reconsidering.

Where it works best: Gemini’s web grounding produces a credible industry benchmark that is hard for older models to source reliably.

Best for: Checking whether your standard NDA template has drifted from what is normal in your specific industry.

3. Recent Case Law / Enforceability Context

You are a research lead checking recent developments that might affect NDA enforceability.
Step 1: Research recent court decisions or regulatory guidance in [jurisdiction] in the last 12-24 months affecting confidentiality agreement enforceability, particularly around [specific concern, e.g. overly broad non-compete-adjacent language, employee NDAs].
Find:
- Relevant recent cases or guidance
- What made a clause unenforceable, where applicable
- Any resulting best-practice recommendations
Step 2: Write a short summary and flag anything in our draft that resembles a pattern found unenforceable.
Our current draft or key terms: [paste them]
Rules: cite sources. This is informational context, not legal advice, and should be confirmed with counsel before relying on it.

Where it works best: Gemini’s web grounding catches recent legal developments that a static knowledge cutoff would simply not know about.

Best for: Higher-stakes NDAs, or any situation where a clause style has started to feel legally aggressive.

4. Comparable Company NDA Practices

You are a research lead benchmarking how comparable companies handle NDAs in a specific situation.
Step 1: Research how companies similar to ours (in size or industry) typically handle [specific situation, e.g. NDAs before a funding round, NDAs with early-stage vendors].
Find:
- Publicly discussed norms or commentary (founder blogs, legal guides, industry publications)
- Common mistakes flagged in this context
- Reasonable defaults commonly recommended
Step 2: Write a short recommendation for our situation based on what you find.
Rules: cite sources. Distinguish between a documented norm and a single opinion piece.

Where it works best: Gemini surfaces practical, real-world commentary that generic legal templates do not capture.

Best for: Early-stage companies or teams without in-house legal counsel who want a sanity check before drafting.

5. Data Privacy Regulation Context For The NDA

You are a research lead making sure an NDA properly accounts for data privacy obligations that sit alongside confidentiality.
Step 1: Research [relevant regulation, e.g. GDPR, CCPA, or the regulation relevant to your jurisdiction] as it applies to [type of data being shared] in an NDA context.
Find:
- Whether confidentiality alone is sufficient, or a separate data processing agreement is typically also required
- Key obligations that should be referenced or cross-referenced in the NDA
Step 2: Write a short note flagging whether our NDA needs an accompanying data processing agreement, and why.
Rules: cite sources. This is informational context, not legal advice.

Where it works best: Gemini catches the common gap where an NDA is treated as sufficient when personal data protection actually requires a separate agreement.

Best for: Any NDA involving personal data, not just general business confidential information.

6. Counterparty Background Check Before Signing

You are a research lead doing basic public diligence before we sign an NDA with a new counterparty.
Step 1: Research [counterparty company name] using public information.
Find:
- Basic company standing (registration, any public litigation history, ownership if publicly known)
- Any public news relevant to reliability as a counterparty
- Whether they operate under a different jurisdiction than stated
Step 2: Write a short summary flagging anything worth confirming before signing.
Rules: only use publicly available information. Cite sources.
This is a basic public-information check, not a substitute for formal due diligence.

Where it works best: Gemini can surface basic public-record context quickly, which is a reasonable first check even though it does not replace formal diligence.

Best for: New counterparties you have not worked with before, especially for higher-stakes disclosures.

Try Proposal.biz for free

How To Get More From Each Prompt

Treat The First Reply As A Draft

Ask for a narrower definition, push back on a term that feels too long, or request a plainer explanation of a specific clause. Each pass sharpens the document before it goes to legal review.

Chain Your Prompts

Once the NDA is signed, the next document is usually a service agreement so keep the relationship’s context handy for whichever document comes next.

Use the output of one prompt as the input to the next. A jurisdiction-research prompt can inform the term clause prompt, which feeds the full-draft prompt.

Save What Works

When a clause or a full NDA works well and clears legal review cleanly, keep it as your standard, with a note on why. Over time you build a reviewed library that is faster than drafting from scratch every time.

From Prompt To Branded Document

AI gives you the words. It does not give you a branded document your counterparty receives cleanly, or tell you when they have opened it. That last stretch, formatting, sending, and knowing whether the other side has actually looked at it, is usually where the time stacks up.

This is where Proposal.biz fits in.

Paste Your Website URL – Proposal.biz pulls your brand assets into a Smart Content Library, so every document you send after the NDA looks like yours automatically.

Generate From A Prompt – Describe what you need and it produces a fully branded document, ready to refine in the Proposal Builder.

Send, Sign And Track – Send a shareable, trackable link instead of a flat PDF, and use built-in e-signing so the NDA gets signed and returned inside the same workflow.

If you would rather start from a ready-made structure instead of a blank page, the NDA template gives you the standard sections already in place, ready to adapt from any of the prompts above.

The simplest workflow: draft your NDA using whichever AI prompt for NDA writing fits the scenario, have it reviewed by counsel, then drop the final copy into Proposal.biz to brand, send and track. You keep the AI tool’s drafting speed and add the document layer that gets it signed, then use any AI prompt to write an NDA you have saved alongside it.

Final Word

An NDA is a small document that carries a lot of weight for how the relationship that follows it feels. Match mutual or one-way to what is actually happening, define confidential information specifically rather than sweepingly, and set a term that fits the sensitivity of the information rather than a default you copied from somewhere else. Every template above produces a strong first draft. None of them replaces a lawyer’s review before signature.

Proposal.biz makes the document side of what comes after simpler. Paste your website URL and your services, case studies and standard terms populate a Smart Content Library you draw from on every document that follows the NDA. The Proposal Builder turns your AI draft into a branded document, a shareable link replaces the PDF, and e-signature keeps the whole chain, NDA through to signed contract, inside one workflow.

Once the NDA clears legal review and the relationship moves to a formal engagement, the prompts in AI prompts for Retainer Agreement cover the ongoing-engagement terms that typically follow.

Try Proposal.biz for free

Frequently Asked Questions

What is the best AI prompt for NDA writing?

There is no single best prompt, and the right one depends on how clear the request is. For a standard mutual or one-way NDA, the ChatGPT templates work well. For a messy internal request or a clause that needs to be precise, the Claude templates handle it most faithfully. For anything involving an unfamiliar jurisdiction or industry norms, the Gemini template researches that context first.

Which AI tool is best for writing an NDA?

Each model has a different strength. ChatGPT is the most flexible for standard mutual and one-way drafts. Claude is best when the request is unclear or a specific clause, like the confidentiality definition or term, needs to be exact. Gemini wins when jurisdiction norms or recent regulatory context should shape the document.

Should an AI-drafted NDA be reviewed by a lawyer before it is signed?

Yes, always. These templates produce a strong, well-structured first draft, but an NDA carries real legal weight and jurisdiction, enforceability and liability questions genuinely need a qualified lawyer’s review before anything goes to a counterparty for signature.

Can I reuse the same NDA prompts across different deals?

Yes. The structure, definition approach and clause logic hold across most business NDAs. What changes is the specific context you feed in: what is actually being disclosed, the jurisdiction, and whether the relationship is mutual or one-directional. Swap those details and the same 18 templates cover most situations.

How do I turn the AI output into a branded, signed document?

AI gives you the words, not a branded document or a way to know whether the counterparty has opened it. Tools like Proposal.biz close that gap: paste your website URL to pull your brand into a Smart Content Library, generate a fully branded document from a prompt, then send a shareable, trackable link and get it e-signed inside the same workflow.