Contract review is one of the most time-intensive tasks in legal practice. AI can speed it up significantly — but only if you understand what it’s good at and where it falls short.

What AI Does Well

Identifying Standard Clauses

AI excels at scanning a contract and flagging whether standard provisions are present or missing: indemnification, limitation of liability, termination, force majeure, governing law, assignment restrictions.

Prompt (ChatGPT):

Review this contract and list all standard commercial contract provisions that are present. Then list any common provisions that are missing. Format as two columns: Present / Missing.

Summarizing Key Terms

For a quick overview of a lengthy agreement:

Prompt:

Summarize the key commercial terms of this contract: parties, term, payment terms, deliverables, termination provisions, liability caps, and any unusual or non-standard clauses. Format as a bullet-point summary.

Comparing to Standards

AI can compare contract language against your firm’s preferred positions.

Prompt:

Compare this indemnification clause to a market-standard mutual indemnification provision. Identify where this clause deviates and whether those deviations favor the buyer or seller. Suggest specific redline language to bring it closer to market standard.

Flagging Risks

Prompt:

Review this contract from the perspective of [party you represent]. Identify the top 5 risk areas, explain why each is a concern, and suggest protective language for each.

What AI Can’t Do

Understand business context. AI doesn’t know your client’s risk tolerance, business relationship with the counterparty, or strategic priorities. A clause that’s unacceptable for one client might be fine for another.

Catch everything. AI misses subtle interactions between clauses — where a broad definition in Section 1 expands liability in Section 8 in ways that aren’t obvious from reading either section alone.

Replace legal judgment. AI can identify that an indemnification clause is one-sided. It can’t tell you whether to push back on it given the deal dynamics.

Guarantee accuracy. AI occasionally misreads contract language, especially complex nested provisions with multiple exceptions and carve-outs.

Dedicated Tools vs ChatGPT

Spellbook works inside Microsoft Word. As you draft or review, it suggests clauses, flags unusual terms, and offers alternatives. It’s trained on legal contracts, so suggestions are more relevant than ChatGPT’s.

Harvey handles complex contract analysis and can process entire deal rooms for due diligence. Better for large-scale review.

CoCounsel integrates with Westlaw for contract analysis backed by legal research.

ChatGPT is free and flexible but has no legal training, no confidentiality guarantees, and no citation to authority. Use it for learning and templates, not for client work.

A Practical Workflow

1. First pass with AI: Upload the contract to your AI tool. Get a summary of key terms and a risk flag report.
2. Human review: Read the AI’s output alongside the actual contract. Verify every flag. Look for what AI missed.
3. Redlining: Use AI to suggest alternative language for problematic clauses. Review and customize each suggestion.
4. Final review: Read the complete contract yourself. AI assists the process — it doesn’t replace it.

Time savings: A contract that takes 2 hours to review manually typically takes 45 minutes with AI assistance. The savings come from the first-pass summary and risk identification, not from skipping human review.

Confidentiality Warning

Before uploading any contract to an AI tool:

• Confirm the tool doesn’t use your data for training
• Check that it’s SOC 2 compliant
• Review your firm’s AI use policy
• Consider whether the contract itself contains confidentiality provisions that restrict sharing with third-party tools

When in doubt, use AI for generic contract templates and frameworks. Add client-specific details manually.

AI makes contract review faster. It doesn’t make it optional. The lawyer’s judgment is still what the client is paying for.