Overview
Each check includes:- Issue definition — What the rule checks for. Example: “Limitation of liability caps direct damages at less than 12 months of fees.”
- Redline guidance — How to mark up the clause when the issue is found. This can include preferred language and fallback positions.
- Comment guidance — What comment to insert explaining the change or flagging the issue for review.
- Notes — Escalation steps, reviewer guidance, or context that stays internal and isn’t shown to counterparties.
Anatomy of a check
A complete rule is composed of three mandatory core elements and several optional parameters:Mandatory elements
- Title — The name displayed in the Word Add-in during an active review.
- Rule — The explicit pass/fail criteria defining what the AI should inspect. If an agreement violates this rule, a “Fail” status triggers your redlining instructions.
- Redline instructions — Specific instructions guiding the AI on how to rewrite, delete, or add language when a contract fails the rule criteria.
Additional features
- Comment guidance & audience — Generates automated contextual remarks explaining why a change was made. You can dictate whether these comments are Internal (for your team only) or External (directed at the counterparty).
- Fallback instructions — Configure up to three alternative concession levels (e.g., “If counterparty rejects a 10-day review period…”). Each fallback includes description of when to use, distinct redline instructions, and comment guidance.
- Scope — Limits execution to First-party only, Third-party only, or Both parties based on whose contract paper is being reviewed.
- Risk score — Sets a severity label (Low, Medium, High, or AI-generated) surfaced when the rule fails.
- Internal documentation — Reference notes for your contract reviewers that do not impact AI decision-making.
Example
Rule: Limitation of liability caps direct damages at less than 12 months of fees. Redline instructions: Revise the cap to 12 months of fees paid in the preceding 12-month period. Comment guidance (external): “We’ve adjusted the liability cap to reflect a standard 12-month fee period.” Fallback: If rejected, propose a cap of 18 months of fees before escalating internally.Build a check
1
Approve AI suggestions
Add checks instantly by reviewing and approving the automated items generated under the AI-suggested checks window.
2
Generate with AI
Under the New Check menu, select Generate with AI, type your requirement (e.g., “Cap liability at 12 months of fees”), and let the platform map out a complete check.
3
Pull from your library
Reuse work by searching your organization’s existing library. You can save an exact linked check or use the Copy feature to create an independent, unlinked duplicate. You can also use pre-configured Pincites Managed industry benchmarks natively provided by Filevine.
4
Build directly from Word
While reviewing a document in Microsoft Word, highlight a text snippet where you applied manual track changes, select the + button, and add context. LOIS for Word uses your real-world redline to construct a playbook check.
Rule writing tips
- Be specific about the trigger. “Indemnification is missing” is vague. “Agreement lacks mutual indemnification for IP infringement claims” is actionable.
- Include fallback positions. Your first ask might get rejected. Give LOIS for Word a second and third option to suggest.
- Match your team’s voice. If your comments are typically collaborative, write comment guidance that sounds collaborative. If you’re more formal, reflect that.
- Add escalation criteria. Use notes to specify when an issue needs senior review or client approval.
Add fallback positions
Fallbacks define acceptable alternative positions within a check, so the AI knows what to suggest when your preferred language is not accepted. You can configure up to three fallbacks, each with its own trigger conditions, redline instructions, and comment guidance. Fallbacks help your team:- Handle different negotiation scenarios, such as small and large contracts.
- Simplify later review passes.
- Apply fallback logic consistently.
- Give the AI enough context to suggest the appropriate position.
1
Open the check
In the Playbook Editor, select the check you want to edit.
2
Add a fallback position
Select Add Fallback, then enter the alternative position and redline instructions.
3
Define when to use it
Add clear conditions, such as a deal value under $50,000, a second-pass review, or the counterparty rejecting your preferred term. Add comment guidance if the fallback should include a comment.
4
Save the check
Select Save. Repeat these steps to add up to three fallback positions.
Example fallback sequence
- Preferred position: Payment is due within 45 days of the invoice date.
- Fallback 1: Allow payment within 60 days for deals under $50,000.
- Fallback 2: Allow payment within 75 days during second-pass negotiations.
Test and refine fallbacks
When testing a check, provide relevant deal context such as “Second pass” or “Deal under $50,000.” Confirm that the expected fallback appears in the review drawer before accepting the edit. Keep each trigger simple and label fallback positions clearly. Review usage over time to identify fallback positions that should become standard.Test your rules
Before deploying a new check across your organization, utilize live testing to verify its accuracy:1
Upload sample agreements
Add sample agreements under the Test Contracts section of your playbook’s Knowledge Base.
2
Open the Test panel
Inside any check editing window, select the Test panel.
3
Run the evaluation
Choose a document and run the evaluation to view a live simulation of the pass/fail result, the analytical explanation, and the exact redlines the AI will execute in production.