Sullivan & Cromwell Builds 'Agreement Analyzer' With OpenAI for Legal Audits
The storied law firm deployed a specialized agent to spot red-flag clauses and risk exposures in complex corporate contracts.
Top corporate law firm Sullivan & Cromwell has teamed up with OpenAI to deploy Agreement Analyzer, an internal AI tool built specifically to comb through multi-hundred-page business contracts and flag non-standard clauses.
Instead of asking an off-the-shelf chatbot to read a messy PDF, the firm integrated legal reasoning playbooks into its workflow. The agent evaluates draft language against decades of firm precedents, identifying ambiguous liability terms, missing warranties, or unusual indemnification clauses that human reviewers might overlook after hours of reading.
Targeted Audits Over Generic Summaries
The tool uses fine-tuning — the practice of taking a pre-trained general AI model and training it further on specialized examples to master a specific industry vocabulary — alongside structured checklists. Rather than writing creative copy, the agent acts like a strict junior associate running a redline comparison.
Agreement Analyzer turns contract review from an exhausting scavenger hunt into an organized risk checklist.
What This Means for You as a Builder
Notice the pattern here: generic 'summarize this document' prompts are being replaced by narrow, extraction-focused agents. When you build workflow tools for finance, HR, or legal teams, your agent should look for specific, predefined fields or risk markers rather than generating fuzzy conversational summaries.