An NDA playbook is the documented set of acceptable, fallback, and walk-away positions an in-house team uses when reviewing or drafting non-disclosure agreements. NDAs are the highest-volume contract most legal teams handle — a typical mid-market in-house team sees 50-300 per quarter — and the playbook is what allows AI tools and paralegals to handle the routine 80% without an attorney reviewing each one.
Mutual vs unilateral
The first and most consequential distinction:
- Mutual NDA. Both parties exchange confidential information. The default for partnerships, vendor evaluations, M&A early-stage discussions, and most B2B engagements where information flows both ways.
- Unilateral NDA. Only one party discloses; only one party has obligations. Common in interview processes, hiring, customer-facing demos, and one-way vendor pitches.
Default to mutual unless there’s a clear reason for unilateral. A unilateral NDA from a counterparty when the engagement is genuinely two-way is a red flag that the counterparty’s legal team isn’t paying attention.
The eight clauses every NDA playbook must address
| Clause | Acceptable | Fallback | Walk-away |
|---|---|---|---|
| Definition of Confidential Information | Marked OR reasonably understood as confidential | Only marked written + identified-as-confidential oral | Only narrowly enumerated categories |
| Term | 2-3 years from disclosure | 5 years | Perpetual obligations |
| Permitted use | Evaluation purpose explicitly stated | Broader business purpose | Any use without limit |
| Carve-outs | Public domain, prior knowledge, independently developed, third-party received | Same plus residuals clause | Missing residuals (in tech contexts) |
| Return / destruction | On request or termination | Within 30 days of termination | No obligation |
| Remedies | Monetary plus injunctive relief | Injunctive only | Liquidated damages without proof |
| Governing law | Mutual jurisdiction (Delaware/NY for US deals) | Counterparty’s home jurisdiction | Foreign jurisdiction without negotiating leverage |
| Assignment | Not without written consent | Permitted to affiliates | Free assignment to any third party |
Industry overlays apply — biotech NDAs add IP non-use clauses, defense-sector NDAs add export-control language, healthcare NDAs add HIPAA acknowledgments.
How to operationalize
- Encode in CLM intake. The intake form asks the requester three questions — counterparty type, dollar value of related deal (if any), industry/jurisdiction — and routes to the right NDA template.
- Auto-approve mutual standard NDAs. LawGeex, Spellbook, or BlackBoiler compares the counterparty’s redlines against the playbook. When all changes fall within “acceptable” tolerance, auto-approve and notify legal.
- Auto-redline counterparty paper. When the counterparty sends their NDA, AI redlines back to the playbook positions automatically. Lawyer reviews the redlined output, not the original document.
- Escalate on three triggers. Walk-away clause appears, novel jurisdiction, or related deal value above the playbook’s auto-approve cap.
- Track cycle time. Mutual standard NDAs should close in under 24 hours. If they’re not, the playbook isn’t being enforced or the AI isn’t trained correctly.
Common pitfalls
- No counterparty playbook intelligence. When the counterparty is a known repeat partner (a Fortune 500 customer, a frequent vendor), a smart playbook stores their typical positions. The negotiation is faster when you know what they’ll insist on.
- Treating NDAs as the most important contract type. They’re not — they’re the most common. The playbook should make NDAs nearly invisible to attorney attention so the team can focus on the 5% of contracts that actually drive material risk.
- Mutual NDA used when unilateral was right. Disclosing party should not voluntarily take on obligations they don’t need. Symmetric forms are convenient but not always correct.
- Perpetual obligations. Some counterparty templates default to “in perpetuity.” Always negotiate to a defined term.
Related
- Contract review SOP — the broader process the NDA playbook fits inside
- MSA redlining rubric — the equivalent for higher-stakes commercial agreements
- Spellbook — Word-native AI that operationalizes the NDA playbook
- LawGeex — auto-approves NDAs that match the playbook