Contract Analysis

The Contract Review Bottleneck

Every commercial transaction starts with a contract. A mid-size law firm might review 500-2,000 contracts per year. A corporate legal department at a Fortune 500 company might handle 10,000+. Each contract must be read, key clauses extracted, risks flagged, and deviations from standard terms identified. This is work that demands precision but involves enormous repetition — the exact pattern where AI excels.

This chapter covers how AI analyzes contracts under US and English common law frameworks, the specific clause types it can extract, and how to build a practical contract review workflow that combines AI speed with attorney judgment.

Contract Types and Legal Frameworks

AI contract analysis tools handle a wide range of commercial agreements. The legal framework governing interpretation varies by jurisdiction and contract type:

Open data/contract-types-global.json in the code panel. This file contains 25+ contract types with their typical clause structures, governing law considerations, and the specific provisions AI tools are trained to extract.

How AI Reads a Contract

Modern contract AI uses a three-layer approach:

Layer 1: Structure Recognition

AI identifies the document's structure — parties, recitals, operative clauses, schedules, and signature blocks. This sounds simple but is critical: a well-structured parse means the AI knows that "Section 8.2(b)" refers to a specific indemnification carve-out, not a general liability provision.

Layer 2: Clause Classification

Each clause is classified by type. Leading tools like Kira Systems can identify over 1,000 distinct clause types. The most commercially important include:

Layer 3: Risk Scoring

AI compares extracted clauses against your organization's playbook — the standard positions your firm or company has pre-approved. Deviations are flagged with severity levels:

Open data/contract-risk-scoring.json to see a detailed risk matrix for the 20 most common clause types, with example language at each risk level.

US-Specific Contract Analysis

The UCC vs Common Law Distinction

A fundamental issue in US contract analysis is whether the Uniform Commercial Code (UCC) or common law governs. AI tools must — and you must verify that they do — correctly identify the governing framework:

When AI reviews a contract, verify that it correctly identifies the governing framework. A clause that is enforceable under common law (e.g., a liquidated damages provision) may be unenforceable under the UCC if it constitutes a penalty.

State-by-State Variation

US contract law varies significantly by state. AI tools must account for:

English Law Contract Analysis

English contract law governs a significant portion of global commercial agreements, particularly in financial services, shipping, insurance, and commodities. Key differences AI must handle:

Building a Contract Review Workflow

Here is a practical workflow combining AI with attorney review:

Total: 45-60 minutes for a standard commercial agreement that would take 2-4 hours manually.

Prompt Engineering for Contract Review

When using general-purpose AI (Claude, ChatGPT) rather than specialized tools, effective prompting is critical:

Prompt: "Review this Master Service Agreement. For each of the
following clause types, extract the relevant language and
assess risk:
1. Limitation of liability (cap amount, excluded damages)
2. Indemnification (scope, carve-outs, cap)
3. Termination (for cause triggers, cure period, convenience)
4. IP ownership (work product, license back, background IP)
5. Governing law and dispute resolution
6. Data protection and confidentiality

For each clause:
- Quote the exact language
- Flag deviations from market-standard terms
- Note any jurisdiction-specific issues (US state law, English
  law, or other)
- Rate risk as LOW / MEDIUM / HIGH with a one-sentence
  explanation"

Key Takeaways