NDA Drafting AI Prompts for Legal Ops

Non-disclosure agreements are the entry point to every significant business relationship. They are also one of the most standardized and, frankly, tedious contract types in legal practice. A well-drafted NDA reflects hours of attorney time; a hastily copied one from a decade-old template can expose your organization to risks that no one bothered to anticipate. The problem is not complexity — the problem is volume. Legal ops teams managing dozens of NDAs at any given time cannot give each one the attention it deserves.

AI changes the equation by handling the structural and drafting work, freeing attorneys to focus on the judgment calls that actually require legal expertise. The goal is not to replace legal review — it is to make that review faster and more consistent by eliminating the mechanical work that introduces errors.

AI Unpacker provides prompts designed to help Legal Ops teams draft NDAs faster, review them more consistently, and build institutional knowledge that improves with every contract cycle.

TL;DR

• AI prompts can generate NDA drafts in minutes that would take attorneys hours to write from scratch.
• The quality of AI-generated NDAs depends entirely on the specificity of the input prompt — generic prompts produce generic contracts.
• Legal ops should build a library of approved NDA clause variations that can be mixed and matched by AI based on counterparty context.
• Reviewing AI-generated drafts requires a structured checklist, not a complete re-draft.
• NDA categorization by counterparty type and transaction context determines which clause variations apply.
• Version control and prompt lineage tracking are essential for audit and compliance.
• AI-generated drafts should always be reviewed by qualified counsel before execution.

Introduction

The average legal ops team handles NDAs as a throughput problem. There are too many, they move too quickly, and the time available per contract is too short. The solution most teams reach for is template standardization: create one good NDA and use it everywhere. This solves the throughput problem but creates a new one — a one-size-fits-all NDA is often either over-engineered (requiring extensive negotiation for routine disclosures) or under-protected (leaving gaps that create risk in material transactions).

The better approach is a modular one: build NDAs from interchangeable clause blocks that can be customized based on the transaction context. This is more work upfront but dramatically reduces negotiation time downstream because both parties are working from a document that was appropriately calibrated to the situation.

AI makes modular drafting practical by generating clause variations on demand, based on structured inputs about the counterparty, the transaction, and the risk profile. Legal ops teams that build effective prompt libraries for NDA drafting find that they can handle 3-5x the NDA volume without proportional increases in attorney time.

This guide provides prompts for four core Legal Ops NDA workflows: intake and categorization, draft generation, review automation, and negotiation support.

1. Intake and Categorization

Before drafting an NDA, you need to know what kind of NDA to draft. The categorization drives the entire drafting process: what information will be disclosed, between whom, in what context, and with what risk appetite. A miscategorized intake produces a misaligned draft that wastes everyone’s time in negotiation.

The Four Variables That Drive NDA Structure

Every NDA is shaped by four categorical decisions: the direction of disclosure (one-way or mutual), the type of parties (corporate, individual, government), the purpose of the relationship (evaluation, employment, M&A due diligence), and the sensitivity of the information (general business, technical/engineering, financial). Getting these four variables right is 80% of the drafting decision.

Prompt for NDA Intake Analysis

Analyze the following NDA intake request and recommend NDA structure.

Intake details:
- Requestor: Business Development Manager
- Counterparty: Tech startup ( Series A, 40 employees, VC-backed)
- Counterparty contact: CEO
- Purpose: Evaluate potential partnership (technology integration)
- Disclosure direction: Not specified by requestor
- Information sensitivity: Technical (API documentation, integration architecture, pricing)
- Urgency: High (counterparty requires NDA before sending documents)
- Our relationship history: First interaction with this counterparty
- Previous NDAs with similar parties: None on file

Tasks:
1. Recommend disclosure direction (one-way or mutual) with rationale:
   - Factors to consider given this counterparty and purpose
   - Risk implications of each direction
2. Identify which NDA template variant to use:
   - Standard (boilerplate protection)
   - Enhanced (for material technology or IP exposure)
   - Streamlined (to accelerate partner evaluation)
3. Flag any red flags from intake:
   - Information sensitivity vs. partnership evaluation scope
   - Urgency pressure (what to verify before expediting)
   - First interaction risk (should this be reviewed by counsel before sending?)
4. Recommend timeline:
   - Standard process (if no red flags)
   - Expedited process (if business need justifies)
   - Points where rushing creates risk

Output format:
- Recommendation with confidence level (High/Medium/Low)
- Key decision points that should be confirmed with requestor
- Draft trigger conditions (what information is needed before drafting can begin)

Prompt for Counterparty Risk Assessment

Not all counterparties present the same risk. A Fortune 500 company is unlikely to misuse your information intentionally; a startup with 6 months of runway may not survive to honor its obligations if it collapses. A counterparty risk assessment adapts your NDA’s protective measures to the actual risk profile.

Conduct a counterparty risk assessment for an NDA with the following counterparty.

Counterparty details:
- Company: Meridian Analytics LLC
- Structure: Limited Liability Company
- Age: 3 years in operation
- Last funding: Seed round, $2M, 18 months ago
- Headcount: 12 employees
- Leadership: 2 co-founders, both previously employed at competitor
- Prior legal actions: None on record
- Reference check: One positive reference from a vendor (partial payment history)
- Physical presence: Remote-first company, no fixed office
- Information requested: Customer lists, pricing models, detailed cost structure

Context: We are considering a potential acquisition. They have requested extensive financial information to conduct due diligence.

Assess:
1. Counterparty legitimacy indicators:
   - Are the stated business activities consistent with the information requested?
   - Does the company structure raise any concerns?
2. Risk factors specific to this NDA type (M&A due diligence):
   - Information sensitivity (highest -- full financials, customer data)
   - Duration risk (due diligence NDAs often run 2-3 years)
   - Exit risk (if deal falls apart, what happens to information?)
3. Specific protective provisions to include:
   - Information return/destruction obligations
   - Permitted disclosure restrictions
   - Non-solicitation clauses (due diligence often reveals talent)
   - Remedy provisions (injunctive relief language)
4. Red flags that warrant legal review before sending:
   - anything that should give us pause before disclosing this information
5. Recommended execution process:
   - Electronic signature acceptable?
   - Who should sign on our behalf?
   - Any approvals required before signing?

Provide a risk rating (Low/Medium/High) with supporting rationale.

2. Draft Generation

Once intake is complete, the drafting phase begins. This is where AI adds the most value — taking structured inputs and generating a complete first draft that an attorney can review, mark up, and approve. The key to effective AI drafting is providing comprehensive context in the prompt, not just a summary of what you need.

Prompt for One-Way NDA (Receiving Party)

Generate a one-way NDA draft for the following scenario.

Scenario:
- Disclosing Party (we): Enterprise software company with proprietary AI/ML technology
- Receiving Party: A potential enterprise customer evaluating our platform
- Purpose: Customer wants to evaluate our platform's fit for their use case, requires access to technical documentation, API specs, and architecture overview
- Disclosure type: Primarily technical (API documentation, white papers, architecture diagrams)
- Term: 3 years
- State/jurisdiction: Delaware

Requirements:
1. Generate a complete NDA including:
   - Parties (full legal names, entity types)
   - Recitals/preamble (purpose statement)
   - Definition of Confidential Information (specific to technical disclosures)
   - Exclusions from Confidential Information (standard carve-outs plus technical-specific)
   - Obligations of Receiving Party (handling, duplication, access restrictions)
   - Permitted Disclosures (legal compulsion, need-to-know employees)
   - Term and Termination
   - Return or Destruction of Information
   - No License or Rights Granted
   - No Warranty (standard)
   - Limitation of Liability (standard)
   - Remedies (injunctive relief availability)
   - Governing Law (Delaware)
   - Entire Agreement
   - Counterparts/Electronic Signatures
   - No Waiver

2. For each section, include:
   - The clause text
   - Alternative language options for different risk profiles
   - Notes on negotiation flexibility (which clauses are market-standard vs. which are client-specific)

3. Flag any terms that may be unfavorable to the Receiving Party (from their perspective, which they may push back on) and suggest alternative language they might propose.

Format the output as a Word-compatible document structure with clear section headings.

Prompt for Mutual NDA (M&A Due Diligence)

Mutual NDAs for M&A due diligence have unique requirements that standard commercial NDAs do not address. The information flows in both directions, the sensitivity is often the highest of any business relationship, and the deal context creates specific risks around insider trading and information barriers.

Generate a mutual NDA for M&A due diligence.

Scenario:
- Buyer: Our company (acquirer)
- Seller: Meridian Analytics LLC (target)
- Purpose: Buyer evaluating potential acquisition of Seller
- Mutual disclosure: Both parties will disclose confidential information
- Information types: Financial statements, customer contracts, intellectual property, employee information, technical product roadmap
- Term: 5 years from execution date
- Governing law: New York

Special provisions to include:
1. Information Barrier/Silent Period provisions:
   - Who within each organization may receive confidential information?
   - How are employees prevented from trading on disclosed information?
   - What happens to information if the deal does not close?
2. Permitted disclosure categories:
   - Legal counsel (with privilege protection)
   - Financial advisors (with NDA coverage)
   - Board and investor notification (limited to need-to-know)
3. Exclusivity provisions (if any):
   - Is this an exclusive or non-exclusive process?
   - What are the consequences of breach?
4. Contact information protocols:
   - Who is the designated point of contact for each party?
   - How is communication routed to prevent inadvertent disclosure?
5. Survival provisions:
   - How long do obligations survive after termination?
   - Specific survival periods for different obligation types
6. Non-solicitation provisions:
   - Are there employee non-solicitation provisions?
   - Are there customer non-solicitation provisions?
   - Standard market terms or enhanced?

Include a risk matrix showing which provisions are market-standard, which favor the Buyer, and which favor the Seller.

Format as a complete legal document with proper numbering and section structure.

3. Review Automation

Reviewing NDA drafts is where legal expertise adds the most value — but only if the review is structured and consistent. AI can help by generating review checklists, flagging non-standard provisions, and comparing incoming drafts against your organization’s approved positions.

Prompt for NDA Review Checklist

Generate a structured NDA review checklist for our legal ops team.

Review context:
- Role: Reviewing NDAs received from counterparties (not our drafts)
- Approver: In-house counsel (non-specialist, supported by AI tools)
- Review time: Maximum 30 minutes per NDA

Checklist structure:
1. Preliminary questions (before reading the NDA):
   - Did we receive this via our standard intake process?
   - Is the purpose consistent with our disclosed business objective?
   - Is the counterparty on our approved vendor list?
2. Key term review:
   - [ ] Disclosure direction (one-way or mutual -- consistent with our request?)
   - [ ] Definition of Confidential Information -- too broad? Too narrow?
   - [ ] Exclusions from Confidential Information -- missing any standard carve-outs?
   - [ ] Term length -- 2-5 years is standard; anything outside this range?
   - [ ] Governing law -- is it a jurisdiction where we are comfortable litigating?
3. Risk provisions:
   - [ ] Limitation of Liability -- is there a cap? Is it reasonable?
   - [ ] Indemnification -- are we indemnifying them for our disclosures?
   - [ ] Remedies -- is injunctive relief available (important for IP)?
   - [ ] No License provision -- ensures we are not granting IP rights inadvertently?
4. Red flags:
   - [ ] Most Favored Nation (MFN) clauses
   - [ ] Non-compete provisions buried in NDA
   - [ ] Automatic renewal provisions
   - [ ] Uncapped liability or unlimited damages provisions
   - [ ] Requirements to notify counterparty of inadvertent disclosure
5. Signature block review:
   - [ ] Is the counterparty signing with full legal authority?
   - [ ] Is our signature authority correct?
   - [ ] Are witnesses or notarization required?

For each checklist item:
- What to look for (specific clause language or structural issue)
- Why it matters (risk or business impact)
- What to do if flagged (accept, negotiate, escalate to specialist)

Prompt for Non-Standard Clause Analysis

Counterparty NDAs frequently include non-standard provisions that deviate from market norms. Identifying these deviations quickly is essential for efficient review. AI can compare incoming clause language against a database of standard provisions and flag deviations for attorney attention.

Analyze the following NDA provision and identify non-standard terms.

Provision (extract from received NDA):

"8. CONFIDENTIAL INFORMATION
8.1 Definition: 'Confidential Information' means any and all information disclosed by either party, whether in written, oral, electronic, visual, or any other form, including but not limited to: business plans, financial data, customer lists, technical specifications, trade secrets, know-how, inventions, and any other information designated as confidential or that reasonably should be understood to be confidential given the nature of the information and circumstances of disclosure.

8.2 Exclusions: Confidential Information shall not include information that: (a) is or becomes publicly available without breach of this Agreement; (b) was rightfully in the possession of the Receiving Party prior to disclosure; (c) is independently developed by the Receiving Party without use of Confidential Information; (d) is rightfully obtained by the Receiving Party from a third party without restriction.

8.3 Additional Exclusions: Confidential Information shall also exclude: (e) information that is required to be disclosed by law, regulation, or court order, provided that the Receiving Party gives prompt written notice to the Disclosing Party prior to disclosure."

Flagged concern by reviewer: "Section 8.3 mentions 'additional exclusions' but the exception for legally compelled disclosure (8.3) requires only 'prompt written notice' -- this is weaker than our standard which requires the Receiving Party to cooperate with the Disclosing Party's efforts to seek protective orders."

Tasks:
1. Identify each non-standard or potentially unfavorable term in this provision
2. For each flagged term:
   - Explain the risk or concern
   - Suggest replacement language that addresses the concern
   - Indicate whether this is a deal-breaker or negotiable
3. Overall assessment:
   - Is this provision closer to standard market terms or non-standard?
   - Priority ranking of items to negotiate (top 3)
   - Items to accept as-is (because counterparty will not budge)

4. Negotiation Support

NDA negotiation is a specialized skill that balances relationship management with risk management. Legal ops teams often handle routine negotiations without attorney involvement, but knowing when to escalate and how to respond to counterparty redlines is critical to both efficiency and risk control.

Prompt for Negotiation Response Generation

Generate negotiation responses for the following counterparty redlines.

Our original NDA (key terms):
- Mutual NDA (both parties are both Disclosing and Receiving)
- Term: 3 years
- Governing Law: Delaware
- Permitted Disclosure: Legal counsel and financial advisors only
- Return of Information: Upon request or termination, Receiving Party must return or destroy all Confidential Information

Counterparty redlines:
1. Change to mutual NDA -> Change to unilateral NDA (Counterparty discloses to us only)
   Rationale: "Our policy is to only sign one-way agreements"
2. Change to Term: 3 years -> 1 year with automatic annual renewal unless either party terminates
   Rationale: "Standard for our company"
3. Change to Governing Law: Delaware -> Counterparty's home state
   Rationale: "More convenient for our legal team"
4. Addition: Permitted Disclosure -> Add "employees and contractors on a need-to-know basis"
   Rationale: "Need flexibility to share with our team"
5. Change to Return of Information -> Remove "upon request" requirement, keep only upon termination
   Rationale: "Too burdensome to respond to ad hoc requests"

Our position (for context):
- We are the party with more confidential information at risk
- We prefer mutual NDAs for due diligence relationships
- We have existing Delaware counsel and prefer Delaware jurisdiction
- We need the ability to request return of information during the relationship (not just at termination)

Generate responses for each redline:
1. Recommended position (Accept / Counter / Escalate to attorney)
2. If Counter: Suggested counter language
3. If Accept: Risk acceptance rationale
4. Talking points for each decision (what to say to counterparty if they push back)

Recommend an overall negotiation strategy and identify any issues that should trigger escalation to in-house counsel.

FAQ

Should AI-generated NDA drafts be reviewed by an attorney?

Yes. AI Unpacker generates drafts that accelerate the drafting process, but they do not replace legal judgment. All AI-generated drafts should be reviewed by qualified counsel before execution.

How do I ensure consistency across AI-generated drafts?

Build a clause library with approved language variations for each standard provision. Use the most specific prompts possible, include your organization’s standard clause language in the prompt, and implement a review checklist that catches deviations from your approved positions.

What is the biggest risk in AI-generated NDAs?

Over-reliance on generic AI outputs that do not reflect your organization’s specific risk appetite or the transaction’s specific context. The remedy is detailed prompts that include counterparty context, transaction type, and risk profile.

Conclusion

NDA drafting is a commodity legal task that benefits enormously from AI assistance. Legal ops teams that build effective prompt libraries and review workflows can handle dramatically more volume without proportional increases in attorney time.

AI Unpacker gives you the prompts to build that library. But the legal judgment — when to protect more, when to concede, when to escalate — that judgment comes from qualified counsel.

Your goal is not to eliminate attorneys from the NDA process. Your goal is to make every attorney hour count for more.