LR-201a · Module 1

What AI Catches — and What It Misses

3 min read

AI is excellent at pattern recognition across large documents. It will catch every instance of uncapped liability language, every circular definition, every provision that contradicts another provision in a different section. It is consistent in ways that human reviewers cannot be — the same risk indicator on the same language every time, without fatigue variance.

AI is poor at context that lives outside the document. The business relationship between the parties. The negotiation dynamic — whether this is a first draft or a fourth redline. The strategic value of the deal that might justify accepting a provision that would normally be flagged. The political reality that the client is also the largest source of referrals. These are judgment calls that require human context, and delegating them to AI is not efficiency — it is abdication.

Do This

  • Use AI for exhaustive clause identification and cross-reference mapping across the full document
  • Rely on AI for consistency — ensuring the same standard is applied to every provision
  • Trust AI pattern matching for known risk indicators: uncapped indemnity, broad IP assignment, silent data retention
  • Validate AI findings against the business context before acting on them

Avoid This

  • Delegate risk tolerance decisions to AI — acceptable risk is a business judgment, not a pattern
  • Trust AI to understand negotiation dynamics or relationship context
  • Assume AI catches everything — novel clause structures that do not match known patterns will be missed
  • Skip human review because the AI flagged nothing — absence of flags is not confirmation of safety