The level of control you exercise over contract language determines how strict or flexible your playbook should be.

There are two archetypes: Prescriptive Playbooks and Principle-Based Playbooks. They serve different functions but work best in combination.
‍

‍Prescriptive Playbooks 🏠 (You control the paper)

When you’re drafting on your own paper and want contracts executed exactly as written, you can be fully prescriptive. You’ve designed the structure, set the defaults, and chosen the language. This gives your AI explicit instructions and eliminates ambiguity.

This approach is ideal when:

  • Your templates represent company policy.
  • Clauses have been legally or commercially approved and must not deviate.
  • Consistency, speed, and risk control outweigh flexibility.

Examples of Prescriptive Rules

  • Exact-match enforcement:
     “Ensure Clause 1.2 uses the following wording exactly: ‘Lorem Ipsum’. Any deviation = reject.”
  • Controlled flexibility:
     “Ensure Clause 1.2 uses this wording, except where the notice period changes. Accept up to 14 days.”

Here, your AI isn’t interpreting intent — it’s enforcing it. It acts as your policy engine, applying precise, clause-referenced logic to every review.
‍

Principle-Based Playbooks 🤝 (You’re reviewing their paper or are not prescriptive about your own)

When you’re reviewing a counterparty’s template or even your own paper but allow flexibility in how clauses are expressed, the rules of engagement shift.

You’re stepping into a space where structure and wording may vary — whether because the drafting originates externally or because your internal policy focuses on outcomes over phrasing.

In these cases, your AI can’t rely on rigid clause references or exact matches; it must rely on principles.

You don’t control the specific drafting — so you control the boundaries of acceptability instead.

Your rules should:

  • Define what outcomes must be achieved rather than how they’re worded.
  • Focus on protections, thresholds, and risk tolerances.
  • Include fallbacks that guide the AI when flexibility is allowed or pushback occurs.

Examples of Principle-Based Rules

  • Outcome-driven check:
     “Ensure the liability cap is at least equal to total fees. If lower, redline to increase.”
  • Fallback-based acceptance:
     “Ensure indemnities are mutual. If unilateral, redline to make them mutual.”

Principle-based playbooks are not about enforcing a script; they’re about preserving intent. They allow linguistic diversity while ensuring the business outcome remains consistent.


Prescriptiveness Framework

Aspect Prescriptive Playbook 🏠 Principle-Based Playbook 🤝
Control You own and enforce exact structure and language. You review external drafting or permit linguistic flexibility on your own.
Rule Style Prescriptive: clause-specific, phrasing-sensitive. Principle-based: outcome-driven, intent-focused.
Fallbacks Tight, explicit, numeric or clause-bound. Tiered, flexible, outcome-based.
AI Behavior Enforce exact matches; reject deviations unless approved. Evaluate protections; redline, accept, or escalate based on thresholds.
Example “Ensure Clause 1.2 matches this wording exactly: ‘Lorem Ipsum’.” “Ensure liability cap ≥ 12 months’ fees. If less, redline; if rejected, fallback to 6 months.”
Back to all posts